PSYC 565 Psychology of Learning
Sweetness668
2012Shorttermandworkingmemoryskillsinprimaryschoolagedchildrenwith.pdf
INT J LANG COMMUN DISORD, JULY–AUGUST 2012, VOL. 47, NO. 4, 457–466
Short Report
Short-term and working memory skills in primary school-aged children with specific language impairment and children with pragmatic language impairment: phonological, linguistic and visuo-spatial aspects
Jenny Freed, Elaine Lockton and Catherine Adams Human Communication and Deafness Group, University of Manchester, Manchester, UK
(Received February 2011; accepted December 2011)
Abstract
Background: Children with specific language impairment (CwSLI) are consistently reported to have short-term memory (STM) and working memory (WM) difficulties. Aim: To compare STM and WM abilities in CwSLI with children with pragmatic language impairment (CwPLI). Methods & Procedures: Primary school-aged CwSLI (n = 12) and CwPLI (n = 23) were assessed on measures of STM and WM. Outcomes & Results: CwPLI had significantly higher scores than CwSLI on STM tasks, and this difference could not be accounted for by differences in receptive language ability. There were no between-group differences for WM tasks. The majority of CwSLI and a significant proportion of CwPLI scored in the impaired range on each of the memory measures. In contrast, some CwPLI scored within the normal range. Conclusions & Implications: CwPLI present with a different profile of STM but not WM ability to CwSLI. Clinical implications are discussed.
Keywords: pragmatic language impairment, specific language impairment, short-term memory, working memory.
What this paper adds The study examined the short-term and working memory skills of children with specific language impairment (CwSLI) and children with pragmatic language impairment (CwPLI). It aimed to replicate existing research relating to the memory skills of CwSLI whilst furthering knowledge of these skills in CwPLI. It is hoped this comparison will develop our understanding of the strengths and weaknesses of memory ability in these children.
Introduction
Children with specific language impairment (CwSLI) and children with pragmatic language impairment (CwPLI) fail to develop language normally despite having normal non-verbal intelligence, normal hearing and possessing no known neurological or emotional difficulties (Rice 2000). Unlike CwSLI, CwPLI have marked difficulty with the interpersonal use of language which is disproportionate to their structural language difficulties (Botting and Conti-Ramsden 1999). CwPLI and CwSLI have similar difficulties with complex language skills such as narrative organization and discourse comprehension (Botting and Adams 2005). Pragmatic features such as topic switching or conversa-
Address correspondence to: Jenny Freed, Human Communication and Deafness, University of Manchester, Oxford Road, Manchester M13 9PL, UK; e-mail: [email protected]
tional dominance, which are typical of CwPLI, are also characteristic of the language of children with high- functioning autism (Tager-Flusberg 2001). It has been suggested that pragmatic language impairment (PLI) may be an intermediate condition between specific language impairment (SLI) and autism (Bishop 2000b).
Short-term memory (STM) tasks measure the ability to hold information in mind for a brief period of time (Vance 2008). Baddeley and Hitch’s (1974) memory model distinguishes between storage of phonological and visuo-spatial information in STM tasks. Research consistently indicates that CwSLI are impaired on tasks of phonological STM such as forwards digit recall and non-word repetition (Archibald and Gathercole 2006;
International Journal of Language & Communication Disorders ISSN 1368-2822 print/ISSN 1460-6984 online c© 2012 Royal College of Speech and Language Therapists
DOI: 10.1111/j.1460-6984.2012.00148.x
458 Jenny Freed et al.
Hick et al. 2005). Impairment in phonological STM is thought to have consequences for vocabulary learning, as phonological representations in STM enable the construction of entries in the lexical store in long-term memory (Baddeley et al. 1998). It has been hypothesized that phonological STM difficulties seen in CwSLI may be explained by poor phonological awareness or phonological representations (Bishop 2000a). This implies that STM impairments in CwSLI might be specific to phonological tasks and would therefore not be expected in visuo-spatial STM tasks, which measure ability to remember images and information about locations.
Most research in this area (e.g. Alloway and Archibald 2008; Archibald and Gathercole 2006) supports the view that CwSLI’s visuo-spatial STM abilities lie within the normal range. However, Hick et al. (2005) did find visuo-spatial STM impairments in their SLI sample and proposed that some CwSLI may have a general STM impairment. Inconsistent findings may be due to the choice of visuo-spatial tasks used. For example, if visual stimuli used in tasks can be stored more effectively by translating into verbal information (e.g., by use of verbal counting strategies), then CwSLI may be at a disadvantage (Johnston 1994).
Not surprisingly, given the presence of phonological STM difficulties, phonological working memory (WM) deficits have also been found in CwSLI on tasks such as backwards digit recall (e.g. Alloway and Archibald 2008; Archibald and Gathercole 2006). WM tasks consist of a storage and processing component and it has been proposed that WM capacity is shared between these two aspects of the task (Just and Carpenter 1992). Therefore, increased processing requirements may limit the amount of information that can be retained. Archibald and Gathercole (2006) suggest that CwSLI may have WM difficulties that are disproportionate to their phonolog- ical STM difficulties and that they may face ‘double memory jeopardy’ (p. 687) with impairments in the storage (phonological loop) and the processing (central executive) components of the WM model (Baddeley and Hitch 1974). According to this theory, it is expected that CwSLI would show impairment on processing aspects of visuo-spatial WM tasks because the central executive is not modality specific. However, visuo-spatial WM skills are generally found to be at an age-appropriate level (e.g. Alloway and Archibald 2008; Archibald and Gathercole 2006) suggesting that CwSLI do not have WM difficulties that are disproportionate to their STM difficulties.
The limited research into the memory skills of CwPLI suggests that they show impairments on phonological STM tasks compared with children with typical language development. Botting and Conti- Ramsden (2003) found that CwPLI performed signifi-
cantly better on a non-word repetition task compared with CwSLI, although all group median scores were in the impaired range compared with children with typical language development. Holck et al. (2009) found that CwPLI performed significantly worse than children with typical language development on a task of forwards digit recall, but attained similar scores to children with typical language development on backwards digit recall.
Children who have pragmatic difficulties may represent a significant proportion of speech–language practitioner caseloads. A survey of service provision in the UK suggested 24.2% of children had some pragmatic problems (Broomfield and Dodd 2004), although this figure includes children with additional learning difficulties. Another study, which focused solely on specialist school placements for children with language disorders in the UK, identified 22% of that cohort as having PLI (Botting and Conti-Ramsden 1999). CwPLI generally require long-term support for learning in the classroom (Botting and Conti-Ramsden 1999). It is therefore important to ascertain information about memory skills in these children as they are likely to play a significant role in learning and in remediation strategies for all children who have language-learning needs. In this preliminary study we aimed to pinpoint specific aspects of memory which are vulnerable in SLI and PLI. This may, in turn, indicate appropriate directions for intervention as well as contributing to knowledge about individual variability within group and continuity of memory skills between groups.
The present study aimed:
• to profile the ability of a group of CwPLI and a group of CwSLI on standardized STM and WM tasks;
• to compare performance between the two clinical groups; and
• to explore whether any between-group differences on STM and WM tasks can be accounted for by receptive language ability.
In the present study, three different types of STM/WM tasks were used to compare the patterns of memory ability between CwSLI and CwPLI. Firstly, ‘phonolog- ical’ tasks, such as digit recall, were those considered by the authors to be less dependent on knowledge of grammar or sentence structure, e.g., digit recall. Secondly, ‘linguistic’ tasks were those tasks that may be supported by general linguistic knowledge, e.g., the sentence recall task. STM is important for sentence recall; however, performance is likely to be influenced by higher-level language skills. Thirdly, tasks described here as ‘visuo-spatial’ were those that require the recall of visual or spatial sequences, for example, cells in a visual matrix. The authors selected visuo-spatial tasks
14606984, 2012, 4, D ow
nloaded from https://onlinelibrary.w
iley.com /doi/10.1111/j.1460-6984.2012.00148.x by L
iberty U niversity, W
iley O nline L
ibrary on [17/04/2023]. See the T erm
s and C onditions (https://onlinelibrary.w
iley.com /term
s-and-conditions) on W iley O
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
icense
Memory skills in pragmatic language impairment 459
they considered least open to phonological encoding, meaning that the stimulus information was less likely to be converted into phonological information for rehearsal.
Method
Participants1
All children were aged between 6;00 and 10;11 years and were on NHS speech-and-language therapy caseloads in the North West of England and South East Scotland. Children attended mainstream schools (or specialist language resources attached to mainstream schools) and were identified as having special educational needs. English was their primary language and speech and language therapists had diagnosed PLI or SLI. In addition, no children had any severe emotional or behaviour needs, speech disorder, severe physical difficulties or hearing loss. All children had non- verbal intelligence abilities within the normal range (≥5th centile) on the Raven’s Coloured Progres- sive Matrices (Raven 1979). This cut-off excluded children with severe non-verbal intelligence difficul- ties, but still left a representative sample of CwPLI and CwSLI.
Children in the PLI group (n = 23, 20 males) were recruited from a cohort of 85 CwPLI taking part in the Social Communication Intervention Project, a randomized controlled trial of intervention (Adams et al. 2012). Additional consent for the current study was obtained from 69 parents, of which 23 children met the following inclusion criteria. CwPLI had a score in the impaired range (≤58) on the General Communication Composite (GCC) and a score ≤8 on the Social Interaction Deviance Composite (SIDC) of the Children’s Communication Checklist—2nd Edition (CCC-2; Bishop 2003a), completed by parents. A score ≤8 on the SIDC was considered to identify children with a disproportionate impairment in pragmatic aspects of communication compared with other aspects of language.
The SLI group (n = 12, eight males) was recruited from NHS speech-and-language therapists in the North West of England and from specialist language resources. CwSLI had a score of <55 on the GCC, and ≥9 on the SIDC of the CCC-2 (Bishop 2003a),2 and either a standard score <8 on the Assessment of Comprehen- sion and Expression naming subtest (ACE; Adams et al. 2001) and/or a standard score <80 on the Test for Reception of Grammar—2nd Edition (TROG-2; Bishop 2003b), i.e. scores in the impaired range. Since SLI is a heterogeneous condition, participants were required to score only in the impaired range on one of the language measures. No language test inclusion
criteria were stipulated for CwPLI since some of this group have normal range structural language abilities. Scores on the GCC of the CCC-2 ranged from 11 to 56 for the PLI group (mean = 29.61, SD = 13.03) and from 15 to 54 for the SLI group (mean = 31.17, SD = 12.25).
All participants’ characteristics can be seen in Table 1, along with the results from t-tests used to check for between-group differences.
Measures
The Automated Working Memory Assessment (AWMA; Alloway 2007) is a computer-administered assessment with subtests measuring STM and WM. Memory tasks were selected to measure phonolog- ical (least reliant on existing linguistic knowledge), linguistic (partly supported by linguistic knowledge), and visuo-spatial (least open to verbal encoding) STM and WM. The AWMA employs a span procedure, whereby the amount of material to be remembered is increased over successive trials. Participants scored a point for each correctly recalled trial. On the WM tasks, the processing and memory aspects of the task were scored separately. The AWMA subtests are detailed below.
Phonological STM
Forwards digit recall [Phonological STM] – the partici- pant hears sequences of digits of increasing length and attempts to recall each sequence in the same order.
Phonological WM
Backwards digit recall [Phonological WM (BDR)] – the participant listens to a sequence of digits and tries to recall the sequence in reverse order.
Counting recall [Phonological WM (CR)] – the participant counts out loud the number of red circles in an array of circles and triangles. At the end of each trial they must recall the number of red circles in each array in the correct order. The partici- pant has to recall a sequence of numbers, so it is similar to the measure of phonological STM (forwards digit recall), but with the addition of a processing task.
Out of the two phonological WM measures, BDR was used in the WM between-group analysis since it was considered to be the purer measure of phonological WM and is widely recognized as a measure of phonological WM. However, as BDR does not have the option of calculating a processing score, CR was used in the WM processing analysis.
14606984, 2012, 4, D ow
nloaded from https://onlinelibrary.w
iley.com /doi/10.1111/j.1460-6984.2012.00148.x by L
iberty U niversity, W
iley O nline L
ibrary on [17/04/2023]. See the T erm
s and C onditions (https://onlinelibrary.w
iley.com /term
s-and-conditions) on W iley O
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
icense
460 Jenny Freed et al.
Table 1. Participant characteristics
PLI group (n = 23, 20 males) SLI group (n = 12, 8 males)
Mean SD Range Mean SD Range Group difference
Age (years;months) 8;3 1;2 6;0−10;4 7;9 1;6 6;2−10;10 p = 0.357 SIDC score on CCC-2 −2.35 8.86 −22 to 8 17.92 8.24 9−31 p < 0.001 Raven’s CPMa 56.57 27.98 17.5−96 46.04 19.26 7.5−82.5 p = 0.201 TROG-2 79.74 15.63 55−109 69.50 9.55 60−95 p = 0.047 ACE Naming 7.91 2.97 3−14 7.33 0.65 6−8 p = 0.378
Notes: aRaw scores on Raven’s Coloured Progressive Matrices are transformed into age-adjusted percentile ranges. For ease of statistical comparison, these ranges have been transformed further into percentile midpoints for that range, e.g. 5th–10th percentile becomes 7.5th percentile. ACE, Assessment of Comprehension and Expression; CCC-2, Children’s Communication Checklist—2; CPM, Coloured Progressive Matrices; SIDC, Social Interaction Deviance Composite; TROG-2, Test for Reception of Grammar.
Linguistic STM
Recalling sentences from the Clinical Evaluation of Language Fundamentals—Fourth UK Edition (CELF-4UK; Semel et al. 2006) [Linguistic STM] – the participant listens to spoken sentences of increasing length and complexity and recalls them verbatim. Each sentence is scored with a maximum score of 3 points, with points deducted for any repetition errors including word omissions, substitutions or transpositions.
Linguistic WM
Listening recall [Linguistic WM] – the participant hears a series of spoken sentences and judges whether each sentence is true or false. At the end of each trial of increasing numbers of sentences, the participant is asked to recall the final word of each sentence in order.
Visuo-spatial STM
Block recall [Visuo-spatial STM] – the participant observes the image of a sequence of blocks being tapped on a computer screen and then attempts to reproduce the sequence in the correct order by tapping the blocks on the screen.
Visuo-spatial WM
Spatial recall [Visuo-spatial WM] – the participant views a picture on the computer screen of two shapes presented next to each other. The participant must first decide whether the shape on the right is the same or opposite to the shape on the left. The shape on the right has a red dot on it which can be in one of three locations. At the end of each trial the participant must recall the location of each red dot on the shape in the correct order.
Procedure
Children completed all testing in a quiet room in their schools and parents were invited into school to complete the CCC-2. The subtests from the AWMA were completed on a laptop computer in one session lasting no longer than 2 h, and regular breaks were given as required. These measures were administered in the order suggested by the AWMA manual. Recalling sentences was completed in a separate assessment session. In accordance with the standardized instruc- tions, children were not told whether they were answering items correctly, but positive encouragement was given throughout the sessions.
Results
Table 2 summarizes standard scores on all of the memory measures. In line with other studies (Alloway et al. 2009; Archibald and Gathercole 2006), it also shows the proportion of children in each group scoring at least 1 SD below the population mean (standard score <85).
Short-term memory – phonological, linguistic and visuo-spatial
Figure 1 shows the distribution of standard scores on the STM tasks (phonological, linguistic and visuo-spatial). One outlier was identified on forwards digit recall in the SLI group. This participant attained a standard score of 96, meaning their performance was in the normal range according to the population norms. There was nothing to suggest that these were invalid data, so this participant remained in the analysis. In order to examine between- group differences, a MANCOVA was conducted with STM type (phonological, linguistic, or visuo-spatial) as the outcome measure and participant group (PLI or SLI) as the between-subjects variable. Receptive language measured by the TROG-2 was included as a covariate. There was homogeneity of covariance matrices (Box’s
14606984, 2012, 4, D ow
nloaded from https://onlinelibrary.w
iley.com /doi/10.1111/j.1460-6984.2012.00148.x by L
iberty U niversity, W
iley O nline L
ibrary on [17/04/2023]. See the T erm
s and C onditions (https://onlinelibrary.w
iley.com /term
s-and-conditions) on W iley O
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
icense
Memory skills in pragmatic language impairment 461
Table 2. Mean, standard deviation range of scores and proportion of children attaining standard scores <85 for the PLI and SLI group on memory measures
PLI group (n = 23) SLI group (n = 12)
Memory type Measure Mean SD Range <85 Mean SD Range <85
Phonological STM Forwards digit recall 87.74 13.90 64−117 0.44 73.83 9.03 64−96 0.92 Linguistic STM Recalling sentencesa 77.17 13.80 55−100 0.61 61.25 6.08 55−70 1.00 Visuo-spatial STM Block recall 96.13 16.06 70−127 0.26 83.58 11.32 66−102 0.58 Phonological WM Counting recall 91.09 14.79 65−111 0.35 85.25 9.46 72−98 0.42 Phonological WM Backwards digit recall 87.70 15.22 58−118 0.35 79.00 15.14 64−109 0.58 Linguistic WM Listening recall 83.21 14.22 62−116 0.48 79.17 11.57 68−99 0.67 Visuo-spatial WM Spatial recall 90.70 12.95 66−118 0.30 90.42 13.35 73−110 0.42
Note: aStandard scores are based on population norms of mean = 10, SD = 3. These scores were converted to standardized scores based on mean = 100, SD = 15, in order to make them directly comparable with the other measures. Standard scores (mean = 100, SD = 15).
Figure 1. Short term memory ability for PLI and SLI groups.
test: mean = 13.89, F (6, 3218.8) = 2.04, p = 0.057). There was a significant effect of group on STM ability (using Pillai’s trace: V = 0.29, F (3, 30) = 4.02, p = 0.016, partial η2 = 0.29), suggesting that the PLI group had significantly higher standard scores than the SLI group and this difference could not be accounted for by the receptive language measure (TROG-2). TROG- 2 was also significant (V = 0.30, F (3, 30) = 4.28, p = 0.012, partial η2 = 0.30). Separate univariate ANCOVAs on the outcome variables (phonological, linguistic and visuo-spatial STM) revealed that the PLI group scored significantly higher than the SLI group on phonological STM (F (1, 32) = 5.97, p = 0.020, partial η2 = 0.16) and linguistic STM (F (1, 32) = 8.65, p = 0.006, partial η2 = 0.21). There was no significant difference between groups on visuo-spatial STM (F (1, 32) = 3.93, p = 0.056, partial η2 = 0.11). Homogene- ity of variances was non-significant for phonological and visuo-spatial STM, but significant (p = 0.01) for linguis-
tic STM, meaning that this particular result should be interpreted with caution.
Working memory – phonological, linguistic and visuo-spatial
Figure 2 shows the distribution of scores on the memory component of the WM tasks (phonological (BDR and CR), linguistic and visuo-spatial). In order to examine between-group differences, a MANCOVA was conducted with WM type (phonological (BDR), linguistic or visuo-spatial) as the outcome measure and participant group (PLI or SLI) as the between-subjects variable. BDR was used for this analysis as the measure of phonological WM (see the Method for justifica- tion). TROG-2 was included as a covariate. There was homogeneity of covariance matrices (Box’s test: mean = 2.20, F (6, 3218.8) = 0.32, p = 0.925). There was no significant effect of group on WM
14606984, 2012, 4, D ow
nloaded from https://onlinelibrary.w
iley.com /doi/10.1111/j.1460-6984.2012.00148.x by L
iberty U niversity, W
iley O nline L
ibrary on [17/04/2023]. See the T erm
s and C onditions (https://onlinelibrary.w
iley.com /term
s-and-conditions) on W iley O
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
icense
462 Jenny Freed et al.
Figure 2. Working memory ability for PLI and SLI groups.
ability (using Pillai’s trace: V = 0.07, F (3, 30) = 0.77, p = 0.518, partial η2 = 0.07) suggesting that WM ability did not differ significantly between groups. TROG-2 was a significant covariate (V = 0.25, F (3, 30) = 3.31, p = 0.033, partial η2 = 0.25). It should be noted that the between-group compari- son was also non-significant if phonological (CR) was included in the analysis instead of phonological (BDR) (p = 0.823).
Processing measures
The WM measures produced a processing score in addition to the memory score. In phonological WM (CR), the processing measure was the child’s ability to count the correct number of red circles (there was no processing measure for phonological WM (BDR), so WM (CR) was used in this analysis). For linguistic WM, this measured the child’s ability to state whether the sentences they heard were true or false. For visuo- spatial WM, this measured the child’s ability to state whether the two shapes were the same or opposite.
Table 3 summarizes the scores on the processing aspect of the WM measures.
One outlier was identified on counting recall processing in the PLI group. This participant attained a standard score of 104, meaning their performance was in the normal range according to the population norms. There was nothing to suggest that these were invalid data, so this participant remained in the analysis.
In order to examine between-group differences, a MANCOVA was conducted with WM processing modality (phonological (CR), linguistic or visuo-spatial) as the outcome measure and participant group (PLI or SLI) as the between-subjects variable. TROG-2 was included as a covariate. There was non-homogeneity of covariance matrices, meaning the results should be interpreted with caution (Box’s test: mean = 17.12, F (6, 3218.8) = 2.52, p = 0.020). There was no signifi- cant effect of group on processing ability (using Pillai’s trace: V = 0.22, F (3, 30) = 2.76, p = 0.059, partial η2 = 0.22) meaning that processing scores did not differ significantly between groups. TROG-2 was a significant covariate (V = 0.27, F (3, 30) = 3.70, p = 0.022, partial η2 = 0.27).
Table 3. Mean, standard deviation, range of scores and proportion of children attaining standard scores <85 on the processing part of the AWMA subtests
PLI group (n = 23) SLI group (n = 12)
Mean SD Range <85 Mean SD Range <85
Counting recall processing 80.61 7.23 71−104 0.83 84.75 5.86 75−96 0.50 Listening recall processing 78.65 6.72 70−91 0.78 80.58 9.01 71−95 0.67 Spatial recall processing 85.43 6.38 72−98 0.30 88.42 10.13 80−109 0.58
Note: Standard scores (mean = 100, SD = 15).
14606984, 2012, 4, D ow
nloaded from https://onlinelibrary.w
iley.com /doi/10.1111/j.1460-6984.2012.00148.x by L
iberty U niversity, W
iley O nline L
ibrary on [17/04/2023]. See the T erm
s and C onditions (https://onlinelibrary.w
iley.com /term
s-and-conditions) on W iley O
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
icense
Memory skills in pragmatic language impairment 463
Figure 3. Memory and processing scores for working memory measures.
Figure 3 shows the distribution of scores for both the memory and processing aspects of the WM tasks split according to participant group. Scores on the memory and processing measures were compared using a series of paired t-tests for each WM task and each partici- pant group separately. On each of the tasks, CwPLI scored higher on the memory aspect compared with the processing aspect (phonological (CR) (t(22) = 3.851, p = 0.001), linguistic (t(22) = 2.259, p = 0.034) and visuo-spatial (t(22) = 2.561, p = 0.018)). For the SLI group there were no significant differences between scores on the memory and processing aspects of the tasks (phonological (CR) (t(11) = 0.187, p = 0.855), linguistic (t(11) = –1.032, p = 0.324) and visuo-spatial (t(11) = 0.715, p = 0.490)).
Discussion
How do CwPLI and CwSLI perform on tests of short-term memory?
Impaired scores (<85 standard score) were observed in the majority (92%) of SLI participants on phonologi- cal STM (forwards digit recall) and all participants on the linguistic STM (recalling sentences) task. Overall, CwSLI scored lowest on the linguistic STM task. These results are consistent with previous studies (e.g. Archibald and Gathercole 2006; Hick et al. 2005) and suggest, not unexpectedly, that CwSLI may not be able to take advantage of linguistic knowledge (e.g. higher- level language skills) to aid recall, unlike their typically developing peers.
CwSLI scored most highly on visuo-spatial STM tasks; however, the mean score for this group was still
lower than 1 SD below the population mean and just over half of the group scored in the impaired range on this task. In this study the authors attempted to select a measure of visuo-spatial STM where the visual informa- tion was less open to being converted into phonological form for rehearsal. The present results suggest that some CwSLI may have a general STM impairment across phonological and visual modalities, which cannot be explained by the disadvantage of being unable to convert visual information into verbal information for rehearsal. This finding is compatible with Hick et al.’s (2005) findings. However, the sample size in this study is small and the contrast to results from other key studies (e.g. Alloway and Archibald 2008; Archibald and Gathercole 2006) requires further exploration.
For CwPLI, the mean score on linguistic STM (recalling sentences) was over 1 SD below population norms; nearly two-thirds of the group had an impaired task score. Approximately half of the PLI group were impaired on phonological STM (forwards digit recall). Compared with the SLI group, there was a much larger range of scores for CwPLI with some children scoring well within the normal range. The vast majority of individual CwPLI or CwSLI obtained their highest standard score on the visuo-spatial task (block recall) and there was individual variability in each group.
Given that CwPLI are reported to possess normal range language skills in terms of phonology (Bishop 2000b), it was predicted that, as a group, these children would score more highly than CwSLI on the measures of phonological and linguistic STM. This was found to be the case, however there was some overlap between groups and a substantial number of children in each group scored within the impaired range. Analysis revealed that
14606984, 2012, 4, D ow
nloaded from https://onlinelibrary.w
iley.com /doi/10.1111/j.1460-6984.2012.00148.x by L
iberty U niversity, W
iley O nline L
ibrary on [17/04/2023]. See the T erm
s and C onditions (https://onlinelibrary.w
iley.com /term
s-and-conditions) on W iley O
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
icense
464 Jenny Freed et al.
this significant group difference could not be accounted for by receptive language ability alone, suggesting that other factors within the two groups’ distinctive language profiles may explain this difference. This is important for our understanding of the nature of PLI and SLI.
How do CwPLI and CwSLI perform on tests of working memory?
Inspection of Table 2 and Figure 2 suggests that, compared with population norms, performance on WM tasks was better than performance on STM tasks for CwSLI with some WM mean scores falling within the normal range. CwSLI presented with a similar pattern of performance across the different modalities of WM tasks as they did across the different types of STM tasks with scores being lowest on the linguistic task (listening recall). However, differences across the different WM tasks were smaller than across the different STM tasks. These results suggest that CwSLI do not have marked WM difficulties beyond those consistent with their poor STM difficulties. The present findings do not support the ‘double memory jeopardy’ hypothesis proposed by Archibald and Gathercole (2006). Inspection of Table 2 shows the PLI group had similar standard scores for STM and WM suggesting a similar level of performance on STM and WM tasks compared with population norms. Between-group analyses found that there was no significant difference between CwPLI and CwSLI on the WM measures (phonological (BDR), linguistic and visuo-spatial WM).
Both groups had task-processing difficulties and the pattern of processing scores reflects the pattern of ability seen in the corresponding memory tasks. Although the AWMA tries to separate out the memory and process- ing aspects of the tasks, each aspect will be affected by the task demands for the other. For example, the ability to complete the processing component will be affected by rehearsal of existing items for the memory component of the task. Rehearsal ability (memory) may also be affected by trying to complete the processing aspect of the task. This interaction between the memory and processing aspects may explain why both groups obtained the highest standard scores on visuo-spatial processing compared with the other modalities since the information could be stored visually. The between- group analysis suggested that the groups showed similar levels of ability on the processing aspects of the task. There were no significant differences between the memory and processing components of the tasks for the SLI group. In contrast, CwPLI obtained higher standard scores on the memory aspect compared with the processing aspect on all three WM tasks. This signifi- cant difference between the two aspects of the task
for the PLI group is interesting and adds to existing knowledge about the difficulties of CwPLI. Further research should pursue the tentative profile of CwPLI as having relative strength in storage (phonological loop) memory but difficulty with the processing components of WM. It would be interesting to investigate this profile in relation to presence of deficits of executive function and central coherence, which have been hypothesized as explanations of pragmatic impairments in CwPLI and in children with high-functioning autism.
Limitations
Although the tasks employed were considered to represent pure measures of phonological, linguistic and visuo-spatial STM and WM, it is possible that higher- level linguistic knowledge may have contributed to the ability on phonological STM tasks and that linguistic ability may have affected performance on visuo-spatial tasks, and vice versa. However, it was important to distinguish between phonological tasks that are more or less reliant on linguistic skills because the present study used a language-impaired population.
Recruitment difficulties resulted in a small SLI sample size, so due caution is noted when interpreting the outcomes. However, the results from the present study are generally in line with previous research examining memory skills in CwSLI. We acknowl- edge that there is a continuum of ability for several aspects of communication between SLI and PLI groups. The diagnostic groupings used, however, allow relevant questions of clinical interest to be researched and are used in similar work in this field. By drawing comparisons between two groups that have some features in common, researchers are better able to tease apart underlying relationships and differences between the disorders (Bishop and Norbury 2005).
Clinical implications and final conclusions
The results suggest that CwPLI do present with a different profile of memory ability compared with CwSLI. In future work it would be interesting to build on this preliminary study by exploring further the heterogeneity of PLI and SLI and the existence of any subgroups.
Whilst the mean scores for some of the measures were within the normal range, the range for each of the measures was large, meaning that a number of children in both groups (but especially the SLI group) demonstrated significant memory problems. Some CwSLI had difficulty with visuo-spatial tasks in addition to difficulties with phonological and linguistic tasks. This implies that some CwSLI cannot necessarily use visual STM strategies to support language learning,
14606984, 2012, 4, D ow
nloaded from https://onlinelibrary.w
iley.com /doi/10.1111/j.1460-6984.2012.00148.x by L
iberty U niversity, W
iley O nline L
ibrary on [17/04/2023]. See the T erm
s and C onditions (https://onlinelibrary.w
iley.com /term
s-and-conditions) on W iley O
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
icense
Memory skills in pragmatic language impairment 465
or they may not translate visual stimuli into phonologi- cal information as easily as their peers. It may be prudent to check visual memory skills of CwSLI as a matter of good practice before implementing visual support for learning strategies in the classroom or therapy. In addition, for those children with memory difficulties, it may be important to reduce memory demands in the classroom and to teach children to use their STM more effectively to facilitate effective learning.
Conversely, the results suggest that there are some children, particularly in the PLI group, who score within the normal range on some aspects of memory despite presenting with a degree of language impairment. These strengths may be utilized during intervention and in the classroom in order to facilitate learning. Once again, it is important to assess different modalities of memory skills in order to obtain a full profile of strengths and difficulties.
Acknowledgements
The authors would like to acknowledge the children, families and all schools involved in the Social Communication Intervention Project for their generous participation; and Kirsty McBean, Jacqueline Gaile, Gillian Earl, Anna Collins and Jenny Gibson for additional testing. This research was supported by a grant to Catherine Adams and Elaine Lockton by the Nuffield Foundation, London, UK. Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.
Notes
1. As all the measures included in the study were standardized, it was not necessary to include children with typical language develop- ment. The use of standardized scores to place group differences into context rather than a control group of children with typical language development is consistent with other studies that have examined memory ability in children with developmental/clinical disorders (e.g. Alloway and Archibald 2008, Alloway et al. 2009, Botting and Conti-Ramsden 2003).
2. The GCC cut-offs are different for the two groups. There are no specific GCC cut-offs for PLI recommended in the manual, so ≤58 was used as this represents a score in the communication impaired range. In contrast, <55 was used for the SLI group as it is the recommended cut-off for SLI in the CCC-2 manual.
References
ADAMS, C., COOKE, R., CRUTCHLEY, A., HESKETH, A. and REEVES, D., 2001, Assessment of Comprehension and Expression 6–11 (Windsor: NFER-Nelson).
ADAMS, C., LOCKTON, E., FREED, J., GAILE, J., EARL, G., MCBEAN, K., et al ., Forthcoming 2012, The Social Communication Intervention Project: a randomised controlled trial of the effectiveness of speech and language therapy for school- age children who have pragmatic and social communication problems with or without autism spectrum disorder. Interna- tional Journal of Language and Communication Disorders.
ALLOWAY, T. P., 2007, Automated Working Memory Assessment (AWMA) (London: Harcourt Assessment).
ALLOWAY, T. P. and ARCHIBALD, L., 2008, Working memory and learning in children with developmental coordina- tion disorder and specific language impairment. Journal of Learning Disabilities, 41(3), 251–262.
ALLOWAY, T. P., RAJENDRAN, G. and ARCHIBALD, L. M. D., 2009, Working memory in children with developmental disorders. Journal of Learning Disabilities, 42(4), 372–382.
ARCHIBALD, L. M. D. and GATHERCOLE, S. E., 2006, Short-term and working memory in specific language impairment. Interna- tional Journal of Language and Communication Disorders, 41(6), 675–693.
BADDELEY, A. D., GATHERCOLE, S. E. and PAPAGNO, C., 1998, The phonological loop as a language learning device. Psychological Review, 105(1), 158–173.
BADDELEY, A. D. and HITCH, G., 1974, Working memory. In G. BOWER (ed.), The Psychology of Learning and Motivation (New York, NY: Academic Press), vol. 8, pp. 47–90.
BISHOP, D. V., 2000a, How does the brain learn language? Insights from the study of children with and without language impairment. Developmental Medicine and Child Neurology, 42(2), 133–142.
BISHOP, D. V., 2000b, Pragmatic language impairment: a correlate of SLI, a distinct subgroup, or part of the autistic continuum? In D. V. BISHOP and L. B. LEONARD (eds), Speech and Language Impairments in Children: Causes, Characteristics, Intervention and Outcome (Hove: Psychology Press), pp. 99–113.
BISHOP, D. V., 2003a, The Children’s Communication Checklist Second Edition (London: Harcourt Assessment/The Psychological Corporation).
BISHOP, D. V., 2003b, Test for Reception of Grammar Second Edition (London: Harcourt).
BISHOP, D. V. and NORBURY, C. F., 2005, Executive functions in children with communication impairments, in relation to autistic symptomatology—I: Generativity. Autism, 9(1), 7– 27.
BOTTING, N. and ADAMS, C., 2005, Semantic and inferencing abilities in children with communication disorders. Interna- tional Journal of Language and Communication Disorders, 40(1), 49–66.
BOTTING, N. and CONTI-RAMSDEN, G., 1999, Pragmatic language impairment without autism: the children in question. Autism, 3(4), 371–396.
BOTTING, N. and CONTI-RAMSDEN, G., 2003, Autism, primary pragmatic difficulties, and specific language impairment: can we distinguish them using psycholinguistic markers? Develop- mental Medicine and Child Neurology, 45(8), 515–524.
BROOMFIELD, J. and DODD, B., 2004, Children with speech and language disability: caseload characteristics. International Journal of Language and Communication Disorders, 39(3), 303–324.
HICK, R., BOTTING, N. and CONTI-RAMSDEN, G., 2005, Short-term memory and vocabulary development in children with Down syndrome and children with specific language impairment. Developmental Medicine and Child Neurology, 47(8), 532– 538.
HOLCK, P., NETTELBLADT, U. and SANDBERG, A. D., 2009, Children with cerebral palsy, spina bifida and pragmatic language impairment: differences and similarities in pragmatic ability. Research in Developmental Disabilities, 30(5), 942–951.
JOHNSTON, J., 1994, Cognitive abilities of children with language impairment. In R. WATKINS and M. L. RICE (eds), Specific Language Impairments in Children (Baltimore, MD: Paul H. Brookes), pp. 107–121.
14606984, 2012, 4, D ow
nloaded from https://onlinelibrary.w
iley.com /doi/10.1111/j.1460-6984.2012.00148.x by L
iberty U niversity, W
iley O nline L
ibrary on [17/04/2023]. See the T erm
s and C onditions (https://onlinelibrary.w
iley.com /term
s-and-conditions) on W iley O
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
icense
466 Jenny Freed et al.
JUST, M. A. and CARPENTER, P. A., 1992, A capacity theory of comprehension—individual-differences in working memory. Psychological Review, 99(1), 122–149.
RAVEN, J. C., 1979, Coloured Progressive Matrices (Oxford: Oxford Psychologists Press).
RICE, M. L., 2000, Grammatical symptoms of specific language impairment. In D. V. BISHOP and L. B. LEONARD (eds), Speech and Language Impairments in Children: Causes, Characteristics, Intervention and Outcome (Hove: Psychology Press), pp. 17– 34.
SEMEL, E., WIIG, E. H. and SECORD, W., 2006, Clinical Evaluation
of Language Fundamentals—Fourth Edition UK (CELF-4UK ) (London: Harcourt Assessment).
TAGER-FLUSBERG, H., 2001, Understanding the language and communicative impairments in autism. In L. M. GLIDDEN
(ed.), International Review of Research in Mental Retardation (San Diego, CA: Academic Press), vol. 23, pp. 185–205.
VANCE, M., 2008, Short-term memory in children with developmen- tal language disorder. In C. F. NORBURY, J. B. TOMBLIN and D. V. BISHOP (eds), Understanding Developmental Language Disorders. From Theory to Practice (Hove: Psychology Press), pp. 23–38.
14606984, 2012, 4, D ow
nloaded from https://onlinelibrary.w
iley.com /doi/10.1111/j.1460-6984.2012.00148.x by L
iberty U niversity, W
iley O nline L
ibrary on [17/04/2023]. See the T erm
s and C onditions (https://onlinelibrary.w
iley.com /term
s-and-conditions) on W iley O
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
icense
