International Journal of Speech-Language Pathology

ISSN: 1754-9507 (Print) 1754-9515 (Online) Journal homepage: www.tandfonline.com/journals/iasl20

Working memory in school-age children with and

without a persistent speech sound disorder

Kelly Farquharson, Tiffany P. Hogan & John E. Bernthal

To cite this article: Kelly Farquharson, Tiffany P. Hogan & John E. Bernthal (2018)
Working memory in school-age children with and without a persistent speech sound
disorder, International Journal of Speech-Language Pathology, 20:4, 422-433, DOI:
10.1080/17549507.2017.1293159

To link to this article: https://doi.org/10.1080/17549507.2017.1293159

View supplementary material

Published online: 17 Mar 2017.

Submit your article to this journal

Article views: 1321

View related articles

View Crossmark data

Citing articles: 6 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=iasl20
International Journal of Speech-Language Pathology, 2018; 20: 422–433

Working memory in school-age children with and without a persistent

speech sound disorder

KELLY FARQUHARSON1 , TIFFANY P. HOGAN2 & JOHN E. BERNTHAL3

1
Department of Communication Sciences and Disorders, Emerson College, Boston, MA, USA, 2MGH-Institute of
Health Professions, Boston, MA, USA, and 3University of Nebraska-Lincoln, Special Education and
Communication Disorders, Lincoln, NE, USA

Abstract
Purpose: The aim of this study was to explore the role of working memory processes as a possible cognitive underpinning of
persistent speech sound disorders (SSD).
Method: Forty school-aged children were enrolled; 20 children with persistent SSD (P-SSD) and 20 typically developing
children. Children participated in three working memory tasks – one to target each of the components in Baddeley’s
working memory model: phonological loop, visual spatial sketchpad and central executive.
Result: Children with P-SSD performed poorly only on the phonological loop tasks compared to their typically developing
age-matched peers. However, mediation analyses revealed that the relation between working memory and a P-SSD was
reliant upon nonverbal intelligence.
Conclusion: These results suggest that co-morbid low-average nonverbal intelligence are linked to poor working memory in
children with P-SSD. Theoretical and clinical implications are discussed.

Keywords: speech sound disorders; working memory; school-aged children; Baddeley’s working memory model; nonverbal
intelligence

Introduction
McCauley, 2000; Munson, Edwards, & Beckman,
Speech sound disorders (SSD) are characterised by 2005; Raine, Hulme, Chadderton, & Bailey, 1991;
delays in the accurate production of age-appropri- Speidel, 1993). Prominent models of working
ate speech sounds (Lewis et al., 2015). As the memory state that it is a cognitive function in
speech sound system develops, children rely on which auditory and/or visual information are tem-
emerging skills in multiple domains including porarily stored and manipulated (Baddeley,
perceptual, cognitive, linguistic, and motoric, for Gathercole, & Papagno, 1998; Cowan, 1988;
accurate speech production. Presumably, an Cowan, Cartwright, Winterowd, & Sherk, 1987;
impairment in any one of these domains could Gathercole & Baddeley, 1990; Gathercole, Willis,
result in abnormal speech sound development. For Baddeley, & Emslie, 1994). Working memory has
approximately 3.9% of children, abnormal speech been shown to significantly contribute to one’s
sound production persists past the age of 8 years ability to perform crucial activities such as reading,
(Lewis et al., 2015; Roulstone, Miller, Wren, & word learning, acquiring language, mathematical
Peters, 2009; Wren, McLeod, White, Miller, & processing and reasoning (Gathercole, Alloway,
Roulstone, 2013). Despite decades of research, the Willis, & Adams, 2006). For speech production, it
causal mechanisms associated with persistent SSD is necessary to accurately and consistently store
(P-SSD) remain elusive (Munson, Baylis, Krause, sounds and readily and appropriately retrieve them
& Yim, 2010). (Oakhill & Kyle, 2000). As such, if this system of
One active area of debate is the impact of working storage and retrieval is deficient, or circumscribed in
memory on the development of speech. Empirical some way, it is likely to manifest as a phonological
data have shown the critical importance of deficit (e.g. speech sound disorder, dyslexia, or
working memory on speech learning in young both). Because phonological skills are necessary for
children (Adams & Gathercole, 1995; Couture & both speech production and literacy success, it is

Correspondence: Kelly Farquharson, Department of Communication Sciences and Disorders, Emerson College, 120 Boylston Street, Boston, MA 02116,
USA. Email: [email protected]
ISSN 1754-9507 print/ISSN 1754-9515 online ß 2017 The Speech Pathology Association of Australia Limited
Published by Informa UK Limited, trading as Taylor & Francis Group
DOI: 10.1080/17549507.2017.1293159
 Working memory and speech sound disorder 423

prudent to examine potential underlying mechan- (1993) found that a child with a history of a speech
isms. Working memory is a plausible contributor, at sound disorder performed more poorly on nonword
least in part, to the speech production deficits seen repetition – a common metric for phonological
in children with SSD. working memory – compared with his typically
developing twin. Follow up studies have reported
Working memory that pre-school aged children (Adams & Gathercole,
1995; Munson, Edwards, & Beckman, 2005) and
There are several conceptualisations of working
school-aged children (Couture & McCauley, 2000;
memory processes. Two prominent models are
Raine, Hulme, Chadderton, & Bailey, 1991) with
offered by Cowan (Cowan, 1988; Cowan,
various SSD also have weaker performance on tasks
Cartwright, Winterowd, & Sherk, 1993) and
that tap phonological working memory, such as
Baddeley (Baddeley & Hitch, 1974) both support
nonword repetition. In addition, Crosbie, Holm,
the overarching role of WM as a ‘‘holding tank’’ for
and Dodd (2009) reported weak executive function-
information while it is either transferred into long-
ing in children with speech sound disorders.
term memory storage, manipulated and immediately
Convergently, there is evidence to support the plaus-
used, or it is forgotten. Additionally, both models
ible role of a phonological working memory deficit in
account for specific components dedicated to the
children with SSD; however, this has yet to be
processing of sensory information. The models
examined in a well-controlled sample of older children
diverge in the mechanisms presumed to process
with persistent SSD. Weak speech production skills
specific sensory information. For instance, Cowan
could make it difficult for children with SSDs to
suggests that the ‘‘active memory’’ processes all
accurately and distinctly reactivate phonological
sensory information; whereas Baddeley hypothesises
information before it decays in memory. In particu-
that there are separate subsystems that process
lar, it is possible that older children who have P-SSD
phonological information versus visual information.
may have inefficient phonological working memory
Thus, for the methodological purpose of the present
skills as a result of many years of inaccurate speech
investigation, it is prudent to explore this cognitive
production. In the present study, we used Baddeley’s
construct within a framework that assumes phono-
Working Memory model (Baddeley & Hitch, 1974)
logical information has a dedicated processing
to examine the three primary components of working
centre. The seminal work of Baddeley and Hitch
memory – phonological loop, visual spatial sketchpad
(1974) proposed a model of working memory with
and central executive – in older children with a
three empirically-supported components: phono-
P-SSD compared to their age-matched typically
logical loop, visual-spatial sketchpad and central
developing peers.
executive. The phonological loop serves as a store
for auditory information, (e.g. speech). The visual–
Working memory in speech sound disorders
spatial sketchpad stores visually presented informa-
tion (e.g. pictures). The primary responsibility of the In a small sample (n ¼ 5) of young school-aged
central executive component is to allocate attention children with SSD, Couture and McCauley (2000)
resources to either the phonological loop or visual– considered the role of working memory within the
spatial sketchpad (Baddeley, 1992). The phono- Baddeley and Hitch (1974) framework. Results
logical loop is described as consisting of two supported weaknesses in phonological memory
subcomponents – a phonological store and an recall, but the authors cautioned the interpretation
articulatory rehearsal mechanism. The phonological due to the small sample size and the unclear direc-
store is a limited capacity space where information is tionality of the phonological memory deficits
organised based on similar features. The articulatory (Couture & McCauley, 2000). An additional limita-
rehearsal mechanism allows information in the tion to the Couture and McCauley (2000) study was
phonological store to be refreshed through subvocal that only the phonological loop was tested; the other
articulation to avoid decay (Baddeley, 2007). It is constructs of the Baddeley and Hitch (1974) model
possible that, for older children with P-SSD, the (i.e. visual spatial sketchpad and central executive)
articulatory rehearsal mechanism is negatively were not examined. Presently, only one investigation
affected by their difficulty producing speech. As included working memory tasks outside the domain
they get older and their speech sound disorder of phonological working memory. Adams and
persists, the lack of maturation of the speech Gathercole (1995) grouped preschoolers as having
production system contributes negatively to their high and low phonological working memory, based
acquisition of strong phonological representations. on their performance on nonword repetition and digit
This can negatively affect a child’s ability to achieve span recall tasks. The children with low phonological
academic success with reading, writing and spelling working memory had more speech errors, but had
(Sutherland & Gillon, 2005). scores similar to children with high phonological
Some evidence suggests that there is a relationship working memory on visual–spatial and central execu-
between phonological working memory and speech tive tasks. Their results suggest that although there
sound production skills. In a case study, Speidel was a link between low phonological working
424 K. Farquharson et al.

memory and speech production skills in pre- 2007); although a divergent report from Raine,
schoolers, there was not a link between visual spatial Hulme, Chadderton, and Bailey (1991) suggested
memory or central executive tasks and speech pro- that short-term (not working) memory impairments
duction skills. Of note, Adams and Gathercole existed in children with SSD in the absence of
(1995) selected participants based on working additional deficits in intelligence. As such, another
memory performance, and not on speech production aim of our study was to examine the influence of
ability. Still, this study lends crucial support to the nonverbal intelligence and expressive vocabulary
possibility that working memory may contribute to on working memory abilities in children with
persistent phonological deficits. In order to expand P-SSD.
upon this work, we present a study that selected
children based on their diagnosis of SSD was
Method
particularly designed to examine the three aspects
of working memory, according to Baddeley and Hitch Participants
(1974), in older children with P-SSD.
Participants were 40 children in second- through
fifth-grade recruited from the Lincoln, Nebraska
The current study
community. Children ranged in age from 7.5 years
In this study, we examined working memory – to 11.8 years (M ¼ 9.3). This age range was selected
phonological loop, visual–spatial sketchpad and because the majority of children should have normal
central executive components of Baddeley’s articulation skills by the age of 8 years (Smit, Hand,
Working Memory model (Baddeley & Hitch, 1974) Freilinger, Bernthal, & Bird, 1990). Thus, if a child
– in school-age children with and without a P-SSD. continues to have difficulty producing speech sounds
Extrapolating from past literature, we predicted that past the age of 8, it is likely a P-SSD. The sample of
children with a P-SSD would perform more poorly 40 was composed of two groups: participants with
than their typically developing peers on a phono- P-SSD (n ¼ 20; 13 males, 7 females) and partici-
logical working memory task while performing pants who were typically developing (n ¼ 20;
similar to their peers on visual–spatial and central 10 males, 10 females). Participants were statistically
executive tasks. Such results would support the matched on age and grade and were recruited as part
notion that children with a P-SSD have working of a larger study of school-age children with a P-SSD
memory limitations specific to the phonological (Farquharson, 2012). A power analysis was con-
loop. Predicated on previous reports, children with ducted to inform sample size, using similar medium
P-SSD are more likely than their typically develop- to large effect sizes as reported in Storkel (2001).
ing peers to have poor expressive vocabulary skills Table I contains descriptive statistics by participant
and low-average nonverbal intelligence (Anthony group. Table II provides details regarding the
et al., 2011; Nathan, Stackhouse, Goulandris, & specific phoneme errors for each child within the
Snowling, 2004; Raitano, Pennington, Tunick, P-SSD group.
Boada, & Shriberg, 2004). Similarly, expressive Children were identified as possible participants
vocabulary skills have been reported to be closely through database searches, local SLP recruitment
tied to nonword repetition skills (Gupta & Tisdale, and flyer distribution. Children with SSD were not
2009; Rvachew & Grawburg, 2008). Both expressive required to be currently receiving treatment.
vocabulary and nonverbal intelligence have been This was primarily because many older children
linked to low working memory performance in with SSD can be dismissed from treatment before
children (Adams & Gathercole, 1995; Baddeley, their disorder is completely resolved (Raitano et al.,

Table I. Descriptive statistics for participants by group, typically developing (TD) and persistent speech sound disorder (SSD).
N Mean SD SEM Min–Max t p
Age in months
SSD 20 112.3 14.93 3.33 90–140
TD 20 113.0 15.26 3.41 89.7–14 0.15 0.879
Grade
SSD 20 3.3 1.1 0.252 2–5
TD 20 3.3 1.1 0.252 2–5 0.00 0.999
Articulationa
SSD 20 80.5 12.9 2.8 52–100
TD 20 104.4 1.9 0.4 101–107 8.16 0.000
Nonverbal IQb
SSD 20 102.1 13.3 2.9 83–130
TD 20 117.3 13.1 2.9 88–136 3.65 0.001
Expressive vocabularyc
P-SSD 20 98.7 15.4 3.4 71–126
TD 20 111 13.2 2.9 89–140 2.7 0.01
a
Goldman–Fristoe Test of Articulation – 2 (Goldman & Fristoe, 2000).
b
Reynolds Intellectual Assessment Scales (Reynolds & Kamphaus, 2003).
c
Expressive One Word Picture Vocabulary Test (Brownell, 2000).
 Working memory and speech sound disorder 425

Table II. GFTA-2 scores and specific phoneme errors. the Expressive One Word Picture Vocabulary Test
Standard Raw Phoneme(s) (Brownell, 2000). The EOWPVT measures naming
Child ID score score Percentile in error ability by showing participants single pictures on
1 75 10 4 /s, z/ pages of an easel and asking the participant to
2 90 5 8 /s, z/ verbally identify the picture (e.g. a picture of a
4 59 23 1 /r, s, z/
8 92 4 7 /r/ windmill; the child should respond that the picture is
10 100 1 26 /r/ a windmill). The groups also differed significantly on
11 100 2 23 /r/ expressive vocabulary (p ¼ 0.01; see Table I). To
14 82 10 48 /r/
15 89 6 9 /r/ examine any relevant contributions of these factors,
16 67 17 4 /r/ we examined the role of nonverbal intelligence and
20 82 10 8 /s, z/ expressive vocabulary in a mediation analysis.
23 74 12 1 /r/
25 78 9 4 /r/
26 63 23 1 /r, s, z/ Children with P-SSD: inclusionary criteria. Children
28 79 10 5 /s, z/ were selected for the P-SSD group based on the
33 93 7 18 /S, y, D/
34 52 28 2 /r, l, k, g/ presence of a speech sound disorder evidenced by a
36 82 9 6 /s, z/ speech sample and a standardised assessment. A
46 91 5 6 /s, z/ speech sample was obtained using a 5–7-min story
48 79 10 5 /r/
52 84 7 6 /s, z/ elicitation task. Each participant was shown a
Mean 80.55 10.4 9.6 wordless picture book and asked to tell the story
SD 12.95 7.18 11.38 that corresponded with the pictures. Speech samples
Min 52 1 1
Max 100 28 48 were transcribed by the first author, an experienced
paediatric speech-language pathologist. Ten per cent
GFTA-2: Goldman-Fristoe Test of Articulation – 2nd Edition.
of speech samples were transcribed by a second
trained listener, a speech-language pathology gradu-
2004) or they no longer qualify for services because ate student. Recall that typically developing children
they have not made adequate progress (Preston & correctly produce all speech sounds in the English
Edwards, 2007). language by 8-years-old (Smit et al., 1990); thus any
consistent speech sound errors were considered to
All participants: inclusionary criteria. All participants be the result of a P-SSD. Articulation skills were also
were monolingual English speakers with normal assessed using a standardised, norm-referenced test
vision (corrected; according to parent question- of single word productions, the Goldman–Fristoe Test
naire), normal hearing and age-appropriate nonver- of Articulation – 2nd Edition (GFTA-2; Goldman &
bal intelligence. To ensure each had normal hearing, Fristoe, 2000). The GFTA-2 elicits articulation of
all participants passed a pure tone hearing screening consonants and consonant clusters in initial, medial,
administered at a level of 25 decibels at 500, 1000, and final word positions and is scored using phonetic
2000 and 4000 Hertz (ASHA, 1997). It was transcription. Participants who had a standard score
anticipated that children with P-SSD would differ less than 85 (1 standard deviation below the mean of
from their peers in both nonverbal intelligence and 100), or a percentile rank below 11, or a consistent
expressive vocabulary (Anthony et al., 2011; Nathan speech sound error(s) during the speech sample
et al., 2004; Raitano et al., 2004). The nonverbal were classified as having a P-SSD. All participants
intelligence subtests from the Reynolds Intellectual classified as having a P-SSD had at least one
Assessment Scales (RIAS, Reynolds & Kamphaus, consistent speech sound error. There were no
2003) were used to confirm age-appropriate non- children in this sample with inconsistent errors. Of
verbal intelligence. Two subtests from the RIAS note, two participants in the P-SSD group achieved
were used to confirm nonverbal intelligence: Odd- standard scores of 100 on the GFTA-2. These two
Item-Out and What’s Missing? The Odd-Item-Out participants (ages 11;0 and 9;6) had a consistent
subtest involved looking at pages from a test book, error on /2/, which is not included as a target sound
each page containing six pictures. Participants were on the GFTA-2. However, six of the words on the
then asked to determine which item ‘‘does not GFTA-2 contain the /2/ phoneme and both of these
belong’’ with the others (e.g. five sedan-style cars participants misarticulated all six productions.
and one large truck – the child should indicate that Additionally, both participants consistently exhib-
the truck does not belong). The What’s Missing? ited this error during conversational speech, which is
subtest involved looking at single pictures and atypical when compared with developmental norms
determining what is missing from the picture (e.g. (Smit et al., 1990). Table II displays the details for
a bird without a beak – the child must identify that each child with a P-SSD.
the beak is missing). Note that, although all partici-
pants had age-appropriate nonverbal intelligence, Children who are typically developing: inclusionary
the P-SSD group was significantly lower than their criteria. Typically developing participants had no
typical peers (p ¼ 0.001; see Table I). Expressive history of speech production difficulties and had
vocabulary was used for descriptive information via never received speech-language treatment, per
426 K. Farquharson et al.

parent report. Normal articulation was validated via accurate serial order. Words that were not recalled
the same speech sample and standardised assess- in serial order were not counted in the PCC or the
ment administered mentioned above. All partici- analyses.
pants in the typically developing group were Great care was taken to control for the speech
required to achieve a standard score above 85 on sound production errors in the P-SSD group. Note
the GFTA-2 and exhibit no articulation errors that there is not presently a standard approach to
(consistent or inconsistent) in the speech sample. scoring such a task, with respect to accounting for a
child’s speech sound errors. Indeed, recent studies
Data collection that have utilised nonword repetition tasks do not
report scoring details (e.g. Lewis et al., 2015;
Data were collected in a quiet and well-lit room in a
Munson, Kurtz, & Windsor, 2005). However,
university clinic, a local school, or a participant’s Dollaghan and Campbell (1998) reported a scoring
home. Each participant received a total of 3 h of data procedure in which phoneme substitutions and
collection administered over the course of 1–2 omissions were counted as incorrect; distortions
months. Each session lasted 60–120 min. Frequent were counted as correct. In the Dollaghan and
breaks were available and taken as necessary/ Campbell (1998) study, the participants were chil-
requested. dren with language impairments, some of which
likely had co-morbid speech sound disorders, but
Experimental tasks that information was not noted (but, see Shriberg
Three experimental tasks were administered to et al., 2009 for an updated syllable repetition task
represent each of the three constructs of working administered to pre-school-aged children).
memory within the Baddeley and Hitch (1974) Similarly, Deevy, Weil, Leonard, and Goffman
model: phonological loop, visual spatial sketchpad (2010) scored a nonword repetition task two differ-
and central executive skills. ent ways: one in which speech sound errors were
counted as response errors and one in which speech
Phonological loop task. Nonword repetition has been sound errors were counted as correct if the phoneme
used extensively in research as a metric of phono- was not in the child’s inventory. In the present
logical working memory skills (see Baddeley, 2003 for investigation, we implemented a specific scoring
review and Gathercole & Baddeley, 1990 for an approach, similar to the second approach in Deevy
example). In the present study, phonological working et al. (2010), to account for the persistent speech
memory was measured by a bespoke serial recall sound errors in school-aged children. For each child,
nonword repetition task administered on a laptop via their consistent pattern of speech errors was
Direct RT software (Jarvis, 2010). The stimuli determined (e.g. /y/ substituted for /s/ in all con-
consisted of consonant–vowel–consonant nonwords texts), and those specific errors were marked as
that followed the phonotatics of the English language, correct. That is, for the child who repeated the
but carried no meaning (see Table III for stimuli list). nonword /hAs/ as /hAy/, the production was scored as
Additionally, although not a focus in the current correct if produced in the correct serial order (e.g.
study, neighbourhood density values were manipu- PCC would be calculated as 100% – two correct
lated for all nonwords (Storkel & Hoover, 2010), consonants produced out of two opportunities).
such that half of the words were dense and half of the This procedure greatly reduced the possibility that
production errors were solely responsible for
words were sparse. Participants listened to rando-
reduced phonological memory in our group with a
mised nonword lists increasing in length: 4 one-
P-SSD. Importantly, we scored the nonword repe-
nonword lists, 4 two-nonword lists, 4 three-nonword
tition task both ways: counting the consistent speech
lists and 4 four-nonword lists. After listening to each
sound error as correct, and also counting it as
list, the participants saw a smiley face on the screen,
incorrect. The P-SSD and TD groups were signifi-
indicating that they were to repeat the nonwords back
cantly different on this task, regardless of scoring
in serial order. Percentage of consonants correct
approach. No children in our sample exhibited
(PCC) was calculated for the words recalled in
inconsistent speech production errors, therefore,
Table III. Nonword repetition stimuli.
any error in production during the nonword repeti-
buk fUS
tion task was assumed to be an error in recall. The
tSæn heb responses to the nonword repetition task were
hæb hAs transcribed using broad transcription. For each
hIb dZep
hIS dZaIf
word, PCC was calculated. An item was counted
kep dZçb as incorrect if it was in the wrong order and/or if it
paIv tAg was recalled with missing or incorrect phonemes
pAb wæb
pAs wAg
(with the exception of the aforementioned speech
tætS jIb sound error correction). Table IV displays the
wUk jub percentage of consonants correct for each group
zæb
at each level. Ten per cent of the sample was
 Working memory and speech sound disorder 427

Table IV. Means, standard deviations, standard error of the mean with P-SSD and their typically developing peers.
and significance tests for NWR task by group.
Because our groups differed significantly on non-
Mean SD SEM p verbal intelligence and expressive vocabulary scores,
NWR 1 word PCC we conducted mediation analyses to determine the
P-SSD 80.55 19.58 3.72 role of these variables in the relation between
TD 81.5 13.06 3.72 0.858
NWR 2 word PCC working memory and the presence of an SSD (see
P-SSD 76.05 15.23 2.98 Supplemental materials for histograms displaying
TD 84.85 11.18 2.98 0.044 the range of expressive vocabulary and nonverbal
NWR 3 word PCC
P-SSD 70.6 16.58 3.6 intelligence scores for children in the P-SSD group).
TD 77.25 15.68 3.6 0.200 Mediation analyses allow for the examination of
NWR 4 word PCC mechanisms that affect the relation between two
P-SSD 50.1 16.18 3.23
TD 63.75 12.5 3.23 0.005 variables (MacKinnon, Fairchild, & Fritz, 2007).
Specifically, mediation analysis examines if a simple
NWR: nonword repetition; P-SSD: persistent speech sound
disorder; TD: typically developing. relation (e.g. the presence of an SSD and working
memory ability) operates via a third variable that is
double-scored and the inter-rater reliability was related to both the predictor and the outcome (e.g.
93%. Any conflicts were resolved between the two nonverbal IQ; Fields, 2013). For our purposes, we
coders until consensus was reached, which resulted wanted to examine if the relation between the
in an ultimate inter-rater reliability of 100% for 10% presence of a P-SSD and working memory ability
of the data that was double-scored. was mediated by either nonverbal intelligence or
expressive vocabulary. If either serve as a mediating
Visual–spatial sketchpad task. Visual–spatial skills variable, it would indicate that there is a change in
were tested using the Spatial Relations subtest from the relation between P-SSDs and working memory
the Woodcock-Johnson III Test of Cognitive Abilities because of the child’s nonverbal intelligence or
(Woodcock, McGrew, & Mater, 2001). This subtest is expressive vocabulary.
reported to measure ‘‘manipulation of visual images in
space’’ in which the participant is required to manipu- Phonological loop task results
late objects ‘‘in the imagination of the ‘mind’s eye’’’ (p.
7). The Spatial Relations task required participants to The nonword repetition task examined phonological
examine four pieces of a puzzle and decide which two working memory by requiring participants to repeat
or three combined to form the intended complete novel phonological sequences of increasing length.
shape. The task increases in complexity as the shapes This task was chosen because it examines serial
are rotated and become more similar. Participants recall (i.e. the ability to repeat a string of phono-
were required to hold visual representations of the logical information in the same order that it was
small shapes in working memory while examining how presented) but eliminates the confounding effects of
those shapes may combine differentially to form the word familiarity that are known to influence per-
larger complete shape. formance (Baddeley, 2003). The nonword repetition
task was scored using per cent of consonants correct
Central executive task. Central Executive function for each word recalled in serial order at each word
was assessed using a stop signal inhibition task length (i.e. 1–4 words). Data were submitted to a
(Gray, Hogan, Alt, Cowan, & Greene, 2011–2016). repeated measures ANOVA with one within-subject
The task was set within a child-friendly ‘‘pirate factor, length (1, 2, 3 or 4 nonwords) and one
game’’ in which monsters invaded the pirate’s island. between-subjects factor, group (P-SSD or typically
Participants saw a variety of monsters appear singu- developing). The results showed a significant main
larly upon the computer screen. Participants were effect of length such that performance decreased
instructed to press the space bar on the computer significantly for all children as the length of non-
each time they saw a monster flash on the screen (i.e. words increased, F (2, 38) ¼ 46.11, p50.001, partial
go trial) unless they heard an auditory signal at the 2 ¼0.548. There was also a significant main effect of
same time they saw the monster (i.e. stop trial). On group, F (2, 38) ¼ 6.347, p ¼ 0.016, partial
the stop trials, the stop signal was presented simul- 2 ¼ 0.143, highlighting that children with a P-SSD
taneously with the visual stimulus. Participants were performed less well than children who were typically
required to inhibit the natural response to press the developing, albeit in the same pattern. There was
space bar. This task is designed to measure inhibi- not a significant interaction between group and
tory control, which is a function of the central nonword length (p ¼ 0.229). Figure 1 illustrates
executive portion of working memory. these results. Follow up t-tests examined group
performance in percentage of consonants correct at
each level of nonword repetition. Table IV displays
Result
these results, indicating that the groups significantly
In this study, we examined performance on three differed when they were to recall two and four
working memory tasks between a group of children nonwords in serial order.
428 K. Farquharson et al.

Figure 1. Group differences on nonword repetition task performance.

To further examine this relation, we employed

mediation analyses (Baron & Kenny, 1986; Hayes,
2013) examining the role of nonverbal intelligence
and expressive vocabulary. Doing so allowed us to
examine the extent to which the relation between
group membership (i.e. P-SSD or typically devel-
opment) and performance on the nonwords repeti-
tion task was changed due to the child’s nonverbal
intelligence score. Our data were submitted to a
mediation regression, with group membership as our
outcome variable, PCC scores from the nonword Figure 2. Results of mediation analysis.
repetition task as the predictor variable, and non-
verbal intelligence as the mediator variable. Results
supported a significant and positive relation between
nonword repetition and nonverbal intelligence, group difference, F (38) ¼ 1.37, p ¼ 0.261, partial
¼ 0.536, p ¼ 0.013. Further, unstandardised indir- 2¼ 0.663. Because this main effect is not signifi-
ect effects were computed using bootstrapping with cant, a mediation analysis is not warranted.
a 95% confidence interval. The confidence interval
did not overlap 0, CI ¼ 0.111 to 0.0074, indicat- Central executive task results
ing significant relation. Figure 2 displays this rela- The stop signal task was developed by Gray et al.
tion. Interestingly, similar results were not found (2011–2016) based on Logan and Cowan (1984).
when expressive vocabulary was the mediating vari- This task examined one aspect of the central
able. Specifically, there was not a significant relation executive – inhibition. Participants were required
between nonword repetition and expressive vocabu- to press a button in response to visual stimuli and to
lary, ¼ 0.050, p ¼ 0.077. The indirect effect of inhibit that response when the visual stimuli were
nonword repetition skills and the presence of a paired with an auditory stimulus. A signal response
P-SSD was not mediated by expressive vocabulary; analysis was used to examine group performance
the 95% confidence interval overlapped 0, (Stanislaw & Todorov, 1999). In doing so, we
CI ¼ 0.115 to 0.0009. computed the average hit rate and average false
alarm rate per group. Hit rate is calculated by
Visual–spatial task results dividing the number of hits (e.g. pressing the space
The visual–spatial task required participants to bar accurately during a go trial) divided by the total
identify which puzzle pieces appropriately paired number of opportunities to hit the space bard
together to form a whole piece. Standard scores accurately (i.e. 54). False alarm rate is calculated
from the spatial relations subtest were used for by dividing the total number of false alarms (i.e.
analysis. Mean standard score for the children with pressing the space bar incorrectly during a stop trial)
P-SSD was 99.35 (SD ¼17.43) and the mean score divided by the total number of stop trials. Next, a z-
for the typically developing children was 106.35 (SD score is computed for the hit rate and the false alarm
¼6.66). Data were submitted to a univariate rate. The z-scored false alarm rate is subtracted from
ANOVA and results revealed a nonsignificant the z-scored hit rate to compute d’. D’ values near 0
 Working memory and speech sound disorder 429

indicate that there was no difference between the go test nonverbal intelligence in children with any form
trials and the stop trials. Larger d’ values indicate a of speech sound disorder. It is commonplace that
difference between the go trials and stop trials. children with linguistic-based disorders perform less
Children in our typically developing group received well than their typically developing peers on non-
a mean d’ score of 4.57 and children in our P-SSD verbal intelligences tests (Lord et al., 1997). As such,
group received a mean d’ score of 5.06. These scores standard practice in studies involving children with
indicate that both groups equally detected a differ- linguistic-based disorders is to include nonverbal
ence between the go trials and the stop trials. intelligence as a co-variate in analyses examining
Because this group difference was not significant, group differences. In the extant literature on speech
mediation analysis was not considered. production skills, nonverbal intelligence is not often
considered. However, an important point of conver-
gence between our work and that of Lewis et al.
Discussion (2015) supports the need to further the consider the
The goal of this study was to examine working influence of nonverbal intelligence in children with P-
memory in school-age children with a P-SSD SSDs.
compared to their typically developing peers. Our A connection between speech production ability
results revealed poor phonological working memory and working memory ability is clear when consider-
for children who had P-SSD. However, the relation ing the role of the articulatory rehearsal mechanism.
between phonological working memory ability and It has been shown that articulation rate contributes
the presence of a P-SSD was mediated by nonverbal to one’s ability to retain information in the phono-
intelligence scores. Interestingly, expressive vocabu- logical loop by keeping phonological information
lary did not mediate this relation, even though ‘‘fresh’’ for recall via the articulatory rehearsal
previous research has reported a strong connection mechanism (Baddeley, 2007). It follows then that
between vocabulary and working memory skills poor articulation ability may contribute to poor
(Gupta & Tisdale, 2009; Rvachew & Grawburg, performance on a phonological loop task. Of course
2008). Two recent studies have corroborated the the inverse is plausible as well. A child with a speech
finding that children with P-SSD often have lower sound disorder would then rely on poorly specified
nonverbal intelligence (Cabbage, Farquharson, & phonological representations to produce sounds in
Hogan, 2015; Lewis et al., 2015). As such, it is words, resulting in the speech sound disorder
interesting to consider the complexities presented by (Sutherland & Gillon, 2005). The role of low
children with P-SSD. We propose three primary average nonverbal intelligence may be explained
plausible explanations for the complexities seen in through the process of redintegration.
the present study; the first of which relate closely to Redintegration refers to a process by which ‘‘. . .lin-
our results and the second two require further guistic knowledge is used to correct errors. . .’’
investigation: (1) deficits in phonological working (Baddeley, 2007, p. 46). Redintegration has been
memory skills, (2) deficits in establishing motoric used to explain why repeating nonwords is more
representations within memory, and (3) deficits in difficult that repeating real words, and why longer
binding between linguistic and motoric skills within words are more difficult to recall than shorter words.
the episodic buffer of working memory. As such, perhaps a child with a speech sound
disorder who has normal or high nonverbal intelli-
Weak phonological working memory gence used redintegration to bootstrap into phono-
logical memory skills. Nonetheless, more work is
In our sample, we found that nonverbal intelligence
needed to disambiguate the complex developmental
mediated the relation between phonological working
and causal associations between poor phonological
memory and the presence of a P-SSD. Although all
working memory, linguistic knowledge linked to
children in the present study had normal nonverbal
redintegration, and poor speech production in chil-
intelligence scores, it appears that children who have
dren with a P-SSD. Additionally, a more robust
a P-SSD paired with a low average nonverbal intel-
examination of not just the phonological loop, but its
ligence are likely to exhibit weak phonological work-
subsidiary components – the phonological store and
ing memory skills. Indeed, similar results were
the articulatory rehearsal mechanism – may provide
reported by Lewis et al. (2015), who longitudinally
insight into the connections between speech pro-
examined outcomes for adolescents who had early
duction and access to items within phonological
speech sound disorders. Specifically, Lewis et al.
working memory.
(2015) found that children between the ages of 11–18
years who had P-SSDs, when compared to children
Deficits in establishing motoric representations within
with and without a history of speech sound disorders,
memory
had weaker working memory skills, evidenced by
performance on a nonwords repetition task, and also Baddeley (2003) suggested that perhaps it is the
had lower nonverbal IQ scores. This is clinically process of setting up speech motor plans that
relevant, considering that it is not commonplace to contribute to the use of the articulatory rehearsal
430 K. Farquharson et al.

mechanism. This, and other models of motor skills and, importantly, provides support for a
production (Guenther, 2006; Levelt, 1983,), may cognitive deficit. Although some researchers have
also explain the findings in the present study. Levelt reported difficulty in determining sensitive ways to
(1983) offers a model of speech production that measure the episodic buffer (Henry, 2010; Nobre
allows for both linguistic and motoric inputs and et al., 2013), it is a logical next step to include in the
outputs. Although the model is comprehensive, it study of this population of children.
does not account for disordered speech production.
Limitation & future direction
Specifically, in his explanation of self-monitoring
and speech repairs (Levelt, 1983), he suggests that Although this study was the first to comprehensively
speakers are always aware of speech errors and stop examine the role of working memory skills in older
to make repairs. However, children with speech children with P-SSD, it is not without limitations.
sound disorders often do not have an awareness of First, although our sample size was determined via an
their errors (Rvachew, Ohberg, Grawburg, & a priori power analysis, it is surely a possibility that we
Heyding, 2003; Rvachew & Jamieson, 1989; may observe different and larger effects with more
Strömbergsson, Wengelin, & House, 2014;). As participants. However, our sample size is in line with
such, this creates a disconnect in the application of the current standard for the study of speech sound
Levelt’s model to children’s with consistent and disorders (Baker & McLeod, 2011). Second, we
persistent speech sound errors. Guenther (2006) in believe future research should include a dynamic,
his Direction into Velocities of Articulators (DIVA) rather than static, assessment of visual spatial skills.
model proposes an auditory and somatosensory Although the task used in the present study is
feedback loop that allows for the development and reported to measure visual spatial manipulation
storage of a motor representation for speech sound skills, it may not be a sensitive measure of visual
production. As such, it is plausible that children with recall. Third, future work should conduct more fine-
P-SSD have a weakness within this feedback loop, grained analyses of responses for serial nonword
which results in aberrant motoric representations. repetition tasks. Such an analysis would provide
Previous research has supported a relation between insights into possible primacy and recency effects that
motor skills and SSD (Krishnan et al., 2013; Peter & may differ between children with P-SSD and their
Stoel-Gammon, 2008; Redle et al., 2015). In such typical peers. Fourth, we did not include a measure of
reports, children with SSD exhibit weaker oral and speech perception abilities. Recent work has reported
fine motor skills compared to typically developing that there are mixed findings regarding the speech
peers. Additionally, children in these studies exhibit perception skills for children with SSD (Cabbage,
low normal scores on language and cognition meas- Farquharson, & Hogan, 2015); however, it has
ures. Interestingly, the analyses in these studies as certainly been implicated as a weakness in young
well as the proposal of the DIVA model does not children with SSD (Rvachew, Ohberg, Grawburg, &
directly account for potential expressive language and Heyding, 2003). Additional studies have suggested
cognitive differences in children with P-SSDs. It is that speech perception is only problematic for a
very likely that motoric skills and linguistic skills subset of children with speech sound disorders
interact in a way that either complements speech (Broen, Strange, Doyle, & Heller, 1983; Rvachew &
production, or works in a negative cycle to attenuate Jamieson, 1989). Finally, we did not collect compre-
speech production skills (Farquharson, 2015; Nip, hensive details regarding the history, severity, and
Green, & Marx, 2009). Certainly, more work is type of speech sound disorder that may have been
needed to validate this relation. present for the children in our sample during their
early years of development. That is, some children
Binding of linguistic and motoric abilities may have exhibited an/r/distortion from early on
whereas other children may have initially had sub-
In 2000 and again in 2012, Baddeley offered updates stantial errors comprised of multiple phonological
to his seminal model from 1974. Both updates processes and only have a residual/r/error at the time
included a fourth component – the episodic buffer. of this study (Karlsson, Shriberg, Flipsen, &
The episodic buffer is conceptualised as a limited McSweeny, 2002; Shriberg, Flipsen, Karlsson, &
capacity space in which information from various McSweeny, 2001). Certainly, this information could
sources is bound together for temporary use or engender different results. Future longitudinal stu-
manipulation (Baddeley, 2000). Presumably, for dies of this nature are necessary to continue to
speech production, the episodic buffer offers a explore the causal mechanisms that contribute to
space to integrate phonological and linguistic repre- protracted speech sound errors, which often cause
sentations with motor representations. As such, it is later literacy and social-emotional deficits.
plausible that children with P-SSD do not have
obvious or radically poor linguistic or motoric skills,
Conclusion
but instead have poor binding. If this is true, this
may explain why children with P-SSDs often have In sum, the results of the present study suggest that
normal, albeit low average, linguistic and motor children with P-SSDs present with complex linguistic
 Working memory and speech sound disorder 431

and cognitive deficits. Our research question References

explored the relation between working memory and Adams, A.M., & Gathercole, S. (1995). Phonological working
the persistence of a speech sound disorder in school- memory and speech production in preschool children. Journal
aged children. Results supported a weakness in of Speech and Hearing Research, 38, 405–414. doi: 10.1044/
phonological working memory in this population of jshr.3802.403
children, when compared to their typically develop- Anthony, J.L., Aghara, R.G., Dunkelberger, M.J., Anthony, T.L.,
Williams, J.M., & Zhang, Z. (2011). What factors place
ing peers. Our follow-up analyses explored the added
children with speech sound disorder at risk for reading
relation of nonverbal intelligence and expressive problems? American Journal of Speech-Language Pathology, 20,
vocabulary. Although expressive vocabulary was not 146–160. doi: 10.1044/1058-0360(2011/10-0053)
related to working memory performance, co-morbid ASHA. (1997). Guidelines for audiologic screening. Rockville, MD:
low-average nonverbal intelligence was linked to poor Author.
working memory in children with P-SSD. Baddeley, A.D. (1992). Working memory. Science, 255, 556–559.
doi: 10.1126/science.1736359
Clinically, these results have substantial import- Baddeley, A. (2000). The episodic buffer: A new component of
ance with respect to assessment. First, it is not working memory? Trends in Cognitive Sciences, 4, 417–423.
commonplace to test nonverbal intelligence (and Baddeley, A. (2003). Working memory: Looking back and looking
sometimes language) in children with speech sound forward. Nature Reviews Neuroscience, 4, 829–839.
disorders. Our results suggest that this construct Baddeley, A.D. (2007). Working memory, thought, and action.
Oxford: Oxford University Press.
may inform prognosis for children with speech Baddeley, A.D., Gathercole, S., & Papagno, C. (1998). The
sound disorders. That is, it is plausible that younger phonological loop as a language learning device. Psychological
children at risk for P-SSDs have low nonverbal Review, 105, 158. doi: 10.1037//0033-295X.105.1.158
intelligence and that this may serve as a red flag for Baddeley, A.D., & Hitch, G.J. (1974). Working memory. The
earlier intervening services. Second, working Psychology of Learning and Motivation, 8, 47–89.
Baddeley, A.D., & Hitch, G.J. (1994). Developments in the
memory appears to be a sensitive measure of
concept of working memory. Neuropsychology, 8, 485. doi:
phonological skills. As such, nonword repetition 10.1037//0894-4105.8.4.485
may be a helpful early screener to determine risk of Baker, E., & McLeod, S. (2011). Evidence-based practice for
persistence in children with SSDs. children with speech sound disorders: Part 1 narrative review.
Language, Speech, and Hearing Services in Schools, 42, 102–139.
doi: 10.1044/0161-1461(2010/09-0075)
Acknowledgements Baron, R., & Kenny, D. (1986). The moderator–mediator variable
distinction in social psychological research: Conceptual, stra-
The authors thank the following funding sources for tegic, and statistical considerations. Journal of Personality and
stimuli and task development as well as participant Social Psychology, 51, 1173–1182. doi: 10.1037//0022-
recruitment support: National Institutes of Health 3514.51.6.1173
R03 Small Grant Award (R03 DC9667) (Hogan) Broen, P.A., Strange, W., Doyle, S.S., & Heller, J.H. (1983).
Perception and production of approximant consonants by
National Institutes of Health R01 (R01 DC010784)
normal and articulation-delayed preschool children. Journal of
(Gray, Alt, Hogan, Cowan, and Green) and Barkley Speech and Hearing Research, 26, 601–608. doi: 10.1044/
Memorial Trust Dissertation Award (Farquharson). jshr.2604.601
The authors thank the following team for the creation Brownell, R. (2000). Expressive one word picture vocabulary test.
of the stop-signal central executive task: Shelley Novato, CA: Academic Therapy Publications.
Gray, Mary Alt, Tiffany Hogan, Nelson Cowan and Cabbage, K.L., Farquharson, K., & Hogan, T.P. (2015,
November). Speech perception and working memory in
Trudy Kuo. The authors thank Jordan Green and Mary
children with residual speech errors: A case study analysis. In
Alt for helpful suggestions on drafts of this manuscript. Seminars in Speech and Language (Vol. 36, No. 234–246).
Germany: Thieme Medical Publishers.
Declaration of interest Couture, A.E., & McCauley, R.J. (2000). Phonological working
memory in children with phonological impairment. Clinical
The authors have no financial interest or benefit that Linguistics & Phonetics, 14, 499–517. doi: 10.1080/
has arisen from the direct applications of this research. 026992000750020332
Cowan, N. (1988). Evolving conceptions of memory storage,
selective attention, and their mutual constraints within the
human information-processing system. Psychological Bulletin,
Supplementary material 104, 163.
Cowan, N. (1993). Activation, attention, and short-term
Supplemental data for this article can be accessed at memory. Memory & Cognition, 21, 162–167. doi: 10.3758/
http://dx.doi.org/10.1080/17549507.2017.1293159 BF03202728
Cowan, N., Cartwright, C., Winterowd, C., & Sherk, M. (1987).
An adult model of preschool children’s speech memory.
Memory & Cognition, 15, 511–517. doi: 10.3758/BF03198385
ORCID Crosbie, S., Holm, A., & Dodd, B. (2009). Cognitive flexibility in
children with and without speech disorder. Child Language
Kelly Farquharson http://orcid.org/0000-0003-
Teaching and Therapy, 25, 250–270. doi: 10.1177/
1836-4681 0265659009102990
Tiffany P. Hogan http://orcid.org/0000-0003-2160- Deevy, P., Weil, L.W., Leonard, L.B., & Goffman, L. (2010).
6995 Extending use of the NRT to preschool-age children with and
432 K. Farquharson et al.

without specific language impairment. Language, Speech, and Analyses of data from the Autism Diagnostic Interview. Journal
Hearing Services in Schools, 41, 277–288. doi: 10.1044/0161- of Autism and Developmental Disorders, 27, 501–517.
1461(2009/08-0096) MacKinnon, D.P., Fairchild, A.J., & Fritz, M.S. (2007).
Dollaghan, C., & Campbell, T.F. (1998). Nonword repetition and Mediation analysis. Annual Review of Psychology, 58, 593. doi:
child language impairment. Journal of Speech, Language, and 10.1146/annurev.psych.58.110405.085542
Hearing Research, 41, 1136–1146. doi: 10.1044/ Munson, B., Edwards, J., & Beckman, M.E. (2005).
jslhr.4105.1136 Relationships between nonword repetition accuracy and
Farquharson, K. (2012). Working memory processes in children with other measures of linguistic development in children with
and without speech sound disorders (Unpublished dissertation). phonological disorders. Journal of Speech, Language, and
University of Nebraska-Lincoln, Lincoln, NE. Hearing Research, 48, 61–78. doi: 10.1044/1092-4388(2005/
Farquharson, K. (2015). Language or motor: Reviewing categor- 006)
ical etiologies of speech sound disorders. Frontiers in Psychology, Munson, B., Baylis, A.L., Krause, M.O., & Yim, D. (2010).
6, 50–59. doi: 10.3389/fpsyg.2015.01708 Representation and access in phonological impairment.
Field, A. (2013). Discovering statistics using SPSS. New York: Sage Laboratory Phonology, 10, 381–404.
Publications. Munson, B., Kurtz, B.A., & Windsor, J. (2005). The influence of
Gathercole, S.E., Alloway, T.P., Willis, C., & Adams, A.M. vocabulary size, phonotactic probability, and wordlikeness on
(2006). Working memory in children with reading disabilities. nonword repetitions of children with and without specific
Journal of Experimental Child Psychology, 93, 265–281. doi: language impairment. Journal of Speech, Language, and Hearing
10.1016/j.jecp.2005.08.003 Research, 48, 1033–1047. doi: 10.1044/1092-4388(2005/072)
Gathercole, S.E., & Baddeley, A.D. (1990). Phonological memory Nathan, L., Stackhouse, J., Goulandris, N., & Snowling, M.J.
deficits in language disordered children: Is there a causal (2004). Educational consequences of developmental speech
connection? Journal of Memory and Language, 29, 336–360. disorder: Keystate 1 National Curriculum assessment results in
doi: 10.1016/0749-596X(90)90004-J English and mathematics. British Journal of Educational
Gathercole, S.E., Willis, C.S., Baddeley, A.D., & Emslie, H. Psychology, 74, 173–186. doi: 10.1348/000709904773839824
(1994). The children’s test of nonword repetition: A test of Nip, I.S.B., Green, J.R., & Marx, D.B. (2009). Early speech
phonological working memory. Memory, 2, 103–127. doi: motor development: Cognitive and linguistic considerations.
10.1080/09658219408258940 Journal of Communication Disorders, 42, 286–298. doi: 10.1016/
Goldman, R., & Fristoe, M. (2000). Goldman–Fristoe test of j.jcomdis.2009.03.008
articulation (2nd ed.). Circle Pines, MN: AGS Publishing. Nobre, A.D.P., Rodrigues, J.D.C., Sbicigo, J.B., Piccolo, L.D.R.,
Gray, S., Hogan, T.P., Alt, M., Cowan, N., & Greene, S. (2011– Zortea, M., Junior, S.D., & de Salles, J.F. (2013). Tasks for
assessment of the episodic buffer: A systematic review.
2016). Working memory and word learning in children with
Psychology & Neuroscience, 6, 331. doi: 10.3922/j.psns.2013.3.10
typical development and language impairment. National
Oakhill, J., & Kyle, F. (2000). The relation between phonological
Institutes of Health R01 (R01 DC010784).
awareness and working memory. Journal of Experimental Child
Guenther, F.H. (2006). Cortical interactions underlying the
Psychology, 75, 152–164. doi: 10.1006/jecp.1999.2529
production of speech sounds. Journal of Communication
Peter, B., & Stoel-Gammon, C. (2008). Central timing deficits in
Disorders, 39, 350–365. doi: 10.1016/j.jcomdis.2006.06.013
subtypes of primary speech disorders. Clinical Linguistics and
Gupta, P., & Tisdale, J. (2009). Does phonological short-term
Phonetics, 22, 171–198. doi: 10.1080/02699200701799825
memory causally determine vocabulary learning? Toward a
Preston, J.L., & Edwards, M.L. (2007). Phonological processing
computational resolution of the debate. Journal of Memory and
skills of adolescents with residual speech sound errors.
Language, 61, 481–502. doi: 10.1016/j.jml.2009.08.001
Language, Speech, and Hearing Services in Schools, 38, 297–
Hayes, A.F. (2013). Introduction to mediation, moderation, and
308. doi: 10.1044/0161-1461(2007/032)
conditional process analysis: A regression-based approach. New
Raine, A., Hulme, C., Chadderton, H., & Bailey, P. (1991).
York: Guilford Press. Verbal short-term memory span in speech-disordered children:
Henry, L.A. (2010). The episodic buffer in children with
Implications for articulatory coding in short-term memory.
intellectual disabilities: An exploratory study. Research in Child Development, 62, 415–423. doi: 10.2307/1131013
Developmental Disabilities, 31, 1609–1614. doi: 10.1016/ Raitano, N.A., Pennington, B.F., Tunick, R.A., Boada, R., &
j.ridd.2010.04.025 Shriberg, L.D. (2004). Pre-literacy skills of subgroups of
Jarvis, B.G. (2010). Direct RT research software (Version 2010). New children with speech sound disorders. Journal of Child
York, NY: Empirisoft. Psychology and Psychiatry, 45, 821–835. doi: 10.1111/j.1469-
Karlsson, H.B., Shriberg, L.D., Flipsen, P., ., & McSweeny, J.L. 7610.2004.00275.x
(2002). Acoustic phenotypes for speech-genetics studies: Redle, E., Vannest, J., Maloney, T., Tsevat, R.K., Eikenberry, S.,
Toward an acoustic marker for residual/s/distortions. Clinical Lewis, B., . . . Holland, S.K. (2015). Functional MRI evidence
Linguistics and Phonetics, 16, 403–424. doi: 10.1080/ for fine motor praxis dysfunction in children with persistent
02699200210128954 speech disorders. Brain Research, 15, 47–56.
Krishnan, S., Alcock, K.J., Mercure, E., Leech, R., Barker, E., Reynolds, C.R., & Kamphaus, R.W. (2003). Reynolds intellectual
Karmiloff-Smith, A., & Dick, F. (2013). Articulating novel assessment scales. Lutz, FL: Psychological Assessment
words: Children’s oromotor skills predict nonword repetition Resources Inc.
abilities. Journal of Speech, Language, and Hearing Research, 56, Roulstone, S., Miller, L.L., Wren, Y., & Peters, T. (2009). The
1800–1812. doi: 10.1044/1092-4388(2013/12-0206) natural history of speech impairment of 8-year-old children in
Levelt, W.J. (1983). Monitoring and self-repair in speech. the Avon Longitudinal Study of Parents and Children: Error
Cognition, 14, 41–104. doi: 10.1016/0010-0277(83)90026-4 rates at 2 and 5 years. International Journal of Speech-Language
Lewis, B.A., Freebairn, L., Tag, J., Ciesla, A.A., Iyengar, S.K., Pathology, 11, 381–391. doi: 10.1080/17549500903125111
Stein, C.M., & Taylor, H.G. (2015). Adolescent outcomes of Rvachew, S., & Grawburg, M. (2008). Reflections on phono-
children with early speech sound disorders with and without logical working memory, letter knowledge, and phonological
language impairment. American Journal of Speech-Language awareness: A reply to Hartmann (2008). Journal of Speech,
Pathology, 24, 150–163. doi: 10.1044/2014_AJSLP-14-0075 Language, and Hearing Research, 51, 1219–1226. doi: 10.1044/
Logan, G. D., & Cowan, W. B. (1984). On the ability to inhibit 1092-4388(2008/07-0158)
thought and action: A theory of an act of control. Psychological Rvachew, S., & Jamieson, D.G. (1989). Perception of voiceless
Review, 91, 295. fricatives by children with a functional articulation disorder.
Lord, C., Pickles, A., McLennan, J., Rutter, M., Bregman, J., Journal of Speech and Hearing Disorders, 54, 193–208. doi:
Folstein, S., . . . Minshew, N. (1997). Diagnosing autism: 10.1044/jshd.5402.193
 Working memory and speech sound disorder 433

Rvachew, S., Ohberg, A., Grawburg, M., & Heyding, J. (2003). Instruments, & Computers, 31, 137–149. doi: 10.3758/
Phonological awareness and phonemic perception in 4-year- BF03207704
old children with delayed expressive phonology skills. American Storkel, H. L. (2001). Learning new words phonotactic prob-
Journal of Speech-Language Pathology, 12, 463–471. doi: ability in language development. Journal of Speech, Language,
10.1044/1058-0360(2003/092) and Hearing Research, 44, 1321–1337.
Shriberg, L.D., Lohmeier, H.L., Campbell, T.F., Dollaghan, Storkel, H.L., & Hoover, J.R. (2010). An online calculator to
C.A., Green, J.R., & Moore, C.A. (2009). A nonword compute phonotactic probability and neighborhood density on
repetition task for speakers with misarticulations: The the basis of child corpora of spoken American English.
Syllable Repetition Task (SRT). Journal of Speech, Language, Behavior Research Methods, 42, 497–506. doi: 10.3758/
and Hearing Research, 52, 1189–1212. doi: 10.1044/1092- BRM.42.2.497
4388(2009/08-0047) Strömbergsson, S., Wengelin, Å., & House, D. (2014). Children’s
Shriberg, L.D., Flipsen, P.J., Karlsson, H.B., & McSweeny, J.L. perception of their synthetically corrected speech production.
(2001). Acoustic phenotypes for speech-genetics studies: An Clinical Linguistics & Phonetics, 28, 373–395.
acoustic marker for residual/¯/distortions. Clinical Linguistics Sutherland & Gillon. (2005). Assessment of phonological repre-
and Phonetics, 15, 631–650. doi: 10.1080/02699200110069429 sentations in children with speech impairment. Language
Smit, A.B., Hand, L., Freilinger, J.J., Bernthal, J.E., & Bird, A. Speech and Hearing Services in Schools, 36, 294–307. doi:
(1990). The Iowa Articulation Norms Project and Its 10.1044/0161-1461(2005/030)
Nebraska Replication. Journal of Speech, Hearing Disorders, Woodcock, R.W., McGrew, K.S., & Mater, N. (2001). Woodcock-
55, 779–798. doi: 10.1044/jshd.5504.779 Johnson III tests of cognitive abilities. Rolling Meadows, IL:
Speidel, G.E. (1993). Phonological short-term memory and Riverside Publishing.
individual differences in learning to speak: A bilingual study. Wren, Y., McLeod, S., White, P., Miller, L.L., & Roulstone, S.
First Language, 13, 69–91. doi: 10.1177/014272379301303705 (2013). Speech characteristics of 8-year-old children: Findings
Stanislaw, H., & Todorov, N. (1999). Calculation of signal from a prospective population study. Journal of Communication
detection theory measures. Behavior Research Methods, Disorders, 46, 53–69. doi: 10.1016/j.jcomdis.2012.08.008