Story retelling and language ability in school-aged children with cerebral palsy and speech impairment
Abstract
Background
Research on retelling ability and cognition is limited in children with cerebral palsy (CP) and speech impairment.
Aims
To explore the impact of expressive and receptive language, narrative discourse dimensions (Narrative Assessment Profile measures), auditory and visual memory, theory of mind (ToM) and non-verbal cognition on the retelling ability of children with CP and speech impairment.
Methods & Procedures
Fifteen speaking children with speech impairment (seven girls, eight boys) (mean age = 11 years, SD = 1;4 years), and different types of CP and different levels of gross motor and cognitive function participated in the present study. Story retelling skills were tested and analysed with the Bus Story Test (BST) and the Narrative Assessment Profile (NAP). Receptive language ability was tested with the Test for Reception of Grammar-2 (TROG-2) and the Peabody Picture Vocabulary Test - IV (PPVT-IV). Non-verbal cognitive level was tested with the Raven's coloured progressive matrices (RCPM), memory functions assessed with the Corsi block-tapping task (CB) and the Digit Span from the Wechsler Intelligence Scale for Children-III. ToM was assessed with the false belief items of the two story tests “Kiki and the Cat” and “Birthday Puppy”.
Outcomes & Results
The children had severe problems with retelling ability corresponding to an age-equivalent of 5;2–6;9 years. Receptive and expressive language, visuo-spatial and auditory memory, non-verbal cognitive level and ToM varied widely within and among the children. Both expressive and receptive language correlated significantly with narrative ability in terms of NAP total scores, so did auditory memory.
Conclusion & Implications
The results suggest that retelling ability in the children with CP in the present study is dependent on language comprehension and production, and memory functions. Consequently, it is important to examine retelling ability together with language and cognitive abilities in these children in order to provide appropriate support.
What this paper adds?
What is already known on this subject?
There is limited research on language ability and cognition in children with CP and different degrees of speech impairment.
What this study adds
The children had problems not only with speech but also with language and cognitive ability. Assessing retelling ability together with receptive language comprehension, non-verbal cognition, memory and ToM gives a greater understanding of how to provide support for children with CP and speech impairment.
Introduction
The United Nations Convention on the Rights of the Child says that ‘Every child has the right to say what they think in all matters affecting them, and to have their views taken seriously’ (UN General Assembly, 1989). Yet, children with cerebral palsy (CP), the most common non-progressive motor disability caused by lesions to the developing brain, are reported also to have speech problems (Pennington et al. 2005, Nordberg et al. 2013, 2014), making it difficult for them to communicate their thoughts. The ability to communicate, typically through speech, is crucial for children in daily life at home and in school. Children with CP constitute a heterogeneous group and, besides problems with speech, other accompanying impairments of varying degree may occur, such as visual deficits and epilepsy (Himmelmann et al. 2010). All this taken together may cause difficulties in communication and social interaction for children with CP.
Narrative ability
One important ingredient in communication is narrative ability. In typically developing (TD) children, this ability develops from the preschool years (Lahey 1988) and continues to be refined throughout childhood and school age. The child needs to be able to master the structure of a story—the beginning, the event and the end—to understand cause–effect relationships and to adapt to the communicative needs of the listener (Lahey 1988, Hudson and Shapiro 1991, Leinonen et al. 2000). Narrative ability provides an excellent quasi-naturalistic measure of children's spontaneous language, a view stated by Botting (2002).
Thus, measuring the specific ability to narrate is considered to be a good way of assessing linguistic, pragmatic and cognitive abilities in children with communication impairments (Norbury and Bishop 2003). In a study of children with late developing language at 2.5 years of age, when followed up at 6 and 8 years it was found that they had persisting difficulties with narrative skills at age 8 (Miniscalco et al. 2007), even though their phonological speech problems were resolved at age 6 (Miniscalco et al. 2005).
Narrative ability can be assessed in various ways, either as story generation (the story is created by the child itself without an adult as a model) or as story retelling. In story retelling, the child retells a story often elicited by pictures. According to Boudreau (2008), story retelling is particularly appropriate for assessing language comprehension in preschool-aged children and suitable for older children with cognitive disabilities. Retold stories have been found to be longer and contain fewer inaccuracies and more grammar components and complete episode structures than generated stories. Another advantage with retold stories is that the examiner is familiar with the content of the story, thus making scoring easier.
Retelling ability is often assessed via the widely used Bus Story Test (BST; Renfrew 1997), which consists of a storybook with 12 pictures and no written words. The narrative is recorded, transcribed orthographically and analysed for (1) the amount of original information included, (2) the number of subordinate clauses and (3) the mean sentence length of the longest five sentences. Low BST scores are a valid predictor of persistent language impairment (Bishop and Edmundsson 1987) and are also associated with literacy disabilities (Stothard et al. 1998, Miniscalco et al. 2007). The BST was used in a study of 55 Swedish-speaking children (6–9 years old) with various neurodevelopmental and communicative impairments (Åhsberg Johnels et al. 2013). Narrative retelling skills, temporal sequencing skills and language skills were measured. It was concluded that language level was important for narrative ability, yet non-verbal sequencing functions were also critical, especially for conveying story information. Narrative skill thus rests on both language and cognitive skills.
Narrative ability in speaking and non-speaking children with CP has not yet been thoroughly investigated. In fact, only one study concerning speaking children with CP has been found (Holck et al. 2011). Story retelling ability was examined in 10 Swedish-speaking children with CP (mean age = 7;11 years) matched for age and gender with 10 TD children using the BST (Renfrew 1997, Svensson and Tuominen-Eriksson 2002). The children had intelligible speech and a diagnosis of spastic diplegia and IQ > 70. Two had language impairment, affecting phonological ability, and minor grammatical difficulties (Holck et al. 2011). No significant differences in information or sentence length between the children with CP and TD were found. However, the children with CP produced significantly fewer subordinate clauses, indicating less grammar complexity.
Earlier research has only focused on such dimensions of narrative discourse as story grammar, without reference to other aspects (e.g. Liles 1987). The Narrative Assessment Profile (NAP) (Bliss et al. 1998) can be used for a further and more thorough analysis of narrative ability. NAP was developed to evaluate the multidimensional nature of discourse with people with communicative impairments. The NAP profile describes the dimensions considered to be crucial for the production of a structurally appropriate narrative. In the study by Holck et al. (2011) on speaking children with CP, the BST analysis was supplemented with NAP analysis (Bliss et al. 1998). The result showed that the children had most problems with the NAP dimension ‘explicitness’ manifested as a shortage of information. It was suggested that the significant relationships found between the explicitness dimension of NAP and the Information score of the BST could be explained by a shortage of information and cohesion.
There are more studies on narrative skills in children with specific language impairment (SLI) than in children with CP. The results have shown that these children experience difficulties in comprehension and production of narratives, compared with TD children (e.g. Bishop and Edmundsson 1987, Merrit and Liles 1989, Botting 2002, Norbury and Bishop 2003). It has been concluded that support of the acquisition of narrative skills is important for this group of children, although additional research is needed (Boudreau 2008).
Language comprehension
The results from research focusing on how speech impairments in children with CP influence the use and development of language have been inconsistent. Smith et al. (2009) found that 16 children with CP (5;5–13 years) without speech and cognitive impairments had significantly lower results than matched speaking controls on a measure of receptive grammar (TROG-2; Bishop 2003). In another study, 12 children without speech and 12 with severe dysarthria (10–18 years) with CP were matched for age and non-verbal cognitive level with 24 TD speaking children with CP (Bishop et al. 1990). The children with speech impairments performed worse than the comparison children on receptive vocabulary, but performed equally well on receptive grammar. Thus, receptive vocabulary seemed to be impaired while grammatical comprehension seemed to be intact.
Memory
Memory is considered one of the basic factors affecting the retelling of stories and should thus be taken into consideration when making conclusions about retelling abilities (Dodwell and Bavin 2008). Memory is complex and central to many cognitive operations. Short-term memory (STM) is the capacity of holding a small amount of information available in mind for a short period of time. Working memory (WM) is the system that actively holds multiple pieces of information in mind, while they or other material can be manipulated (Baddeley 1986). The limited duration of STM suggests that its contents decay over time (Baddeley and Hitch 1974). Only a few studies on memory in speaking or non-speaking children with CP have been presented. In one study, 15 non-speaking children with CP (5–12 years) were matched with typically speaking controls (Larsson and Dahlgren Sandberg 2008). The results showed that the children had problems with auditory short-term memory and WM measured with digit span (Wechsler 1999). Furthermore they had difficulties with visuo-spatial memory measured with the Corsi block-tapping test (Milner 1971). These findings were seen as support for the hypothesis that lack of articulatory ability would indicate weaker memory function in individuals with severe speech difficulties (Baddeley 1986). Bishop and Robson (1989) examined STM in 24 children with CP and speech impairment (anarthric/dysarthric) and compared them with 24 children with CP without speech impairment. The memory span was assessed by letting the children eye-point at series of pictures of increasing length, placed on a board in front of them. No impairment of memory span was found in the children with CP and speech impairment compared with the controls without speech impairment. The results challenge the notion that motoric ‘articulatory coding’ is implicated in STM (Baddeley 1986) and suggest that processes such as rehearsal and phonemic segmentation involve generation of a more abstract central articulatory code. The developmental path of memory skills in children with speech impairment, in comparison with that in TD children, is not known from the study by Bishop and Robson (1989). These considerations influenced Murray and Goldbart (2011) to investigate the emergence of memory strategies in 30 children with CP (3–6 years) in comparison with that of age-matched TD peers. The same procedure as in the original Bishop and Robson (1989) study was used. They found that articulatory rehearsal of nouns, as a memory strategy, developmentally seemed to precede adjective rehearsal. Adjective rehearsal in turn preceded verb rehearsal in both the children with CP and the TD children and input modality did not influence rehearsal of either verbs or adjectives. A qualitative difference in memory performance was found with a broader range of memory strategies in the TD group (Murray and Goldbart 2011).
Baddeley et al. (1975) have suggested that articulation rehearsal rate determines the amount of verbal material that can be maintained in memory. Interestingly, White et al. (1994) demonstrated that normal speech rates did not play a crucial role in determining the capacity of memory span. Eleven children with spastic diplegic CP without cognitive disabilities (5–11 years) and with subtle impairments in articulatory rate were compared with typically speaking controls. They did not show any impairment of memory span compared with the children without speech rate deficits when tested with a memory span task consisting of verbal repetition of one-, two- or three-syllable words. In accordance with these contradictory results, memory problems need to be further investigated in more children with and without speech.
Sometimes children's speech does not meet all their communication needs and they may rely on augmentative and alternative communication (AAC) strategies, such as visual characters or computers with voice output. Thistle and Wilkinson (2013) have emphasized that AAC systems may have some unique WM demands for the AAC users. It is important to examine how WM might interact with AAC use and the role of the WM demands of AAC. Examination of memory functions in speaking individuals with typical development as well as in AAC users with speech impairment are necessary in order to identify if there are any differences between the WM demands for different groups of people.
Theory of mind (ToM)
ToM is another important and complex ability where language and cognition are considered to interact (Miller 2004). ToM is important for narration and can be defined as ‘the ability to predict and explain people's behaviour with reference to mental state’ (Slaughter and Repacholi 2003:1). Carlson et al. (2013) described ToM as an important ability affecting many aspects of child development including social skills and early school success.
A task often used to determine if a child has acquired a ToM is the so-called ‘false belief’ (FB) task, a test of the child's ability to understand that beliefs also can be false. It is not enough merely to observe the situation to predict the other person's behaviour in an FB task, the child also needs the ability to put him/herself into the other person's world of thoughts (Rakhlin et al. 2011). First-order FB tasks involve the ability to realize that it is possible to hold an FB with regard to real events in the world. The more advanced second-order belief tasks are related to the ability to understand an FB of one person based on the thoughts of another person. This capacity to understand not only one person's perception (first-order reasoning) of a situation but also different individuals’ concern about other persons’ mental states (second-order reasoning) underlies much of our social reasoning and is important for all sophisticated understanding of human behaviour.
Poor performance on ToM was found in 14 non-speaking children (mean age = 9 years) with CP and non-verbal cognitive disabilities and the results were discussed in terms of poor language abilities (Dahlgren et al. 2010). In a study by Holck et al. (2010), 10 speaking children with CP (mean age = 7;11 years) performed at a lower level than TD children matched for chronological age (CA) and gender on the FB condition of the so-called thought picture test (ToM test) (Woolfe et al. 2002). The two studies by Holck et al., one on ToM (Holck et al. 2010) and the other on narrative ability (Holck et al. 2011), cited above, concern the same study samples.
- expressive language ability (in terms of the retelling measures, sentence length and subordinate clauses);
- narrative discourse dimensions (NAP measures);
- receptive language ability (in terms of the measures, receptive grammatical understanding and receptive vocabulary);
- auditory and visual memory; and
- ToM and non-verbal cognition.
Method
Participants
The children with CP and speech impairment in this study were recruited from a population-based study of 129 children (66 girls, 63 boys), born in 1999–2002, which forms part of the panorama of CP in the western Sweden study area (Himmelmann et al. 2010). Twenty-seven of the original 129 children of the study had speech impairment documented in the medical and speech and language therapy records (Himmelmann et al. 2010; Nordberg et al. 2013). Twenty-two of the 27 children participated, with the parents’ consent, to be part of an overall assessment of speech, language and communication skills, made by the first author (Nordberg et al. 2013, 2014).
Seven of the 22 children from the population-based study were excluded from the present study. One child was not able to take part in the assessment due to epilepsy; five children were not able to produce five spoken utterances, which is necessary for the scoring procedure of the retelling assessment with the BST (Renfrew 1997); and one child did not participate due to visual deficits (the retelling test is picture based). Consequently, the study group comprised of 15 Swedish-speaking children with CP (seven girls, eight boys), CA 9;2–12;9 (mean = 11;0 years, SD = 1;4 years). All had intelligible speech. The characteristics of the 15 children are shown in table 1. Two children (Child C [Assyrian] and Child J, [Albanian]) were bilingual, with Swedish at the same level as their native tongue according to their teacher.
| Child | Age | Type of | Gross-motor | |
|---|---|---|---|---|
| (N = 15) | Gender | (years;months) | cerebral palsy | function level |
| A | Female | 9;6 | USCP | I |
| B | Male | 9;9 | USCP | I |
| C | Male | 12;8 | USCPa | I |
| D | Male | 11;11 | USCP | I |
| E | Female | 12;9 | USCP | I |
| F | Female | 12;8 | USCP | I |
| G | Female | 9;8 | BSCPb | IV |
| H | Male | 10;7 | BSCP | IV |
| I | Male | 11;3 | USCP | III |
| J | Male | 11;8 | USCP | I |
| K | Male | 11;10 | Dyskinetic | III |
| L | Female | 9;2 | Dyskinetic | II |
| M | Female | 10;1 | Ataxic | I |
| N | Female | 9;2 | Ataxic | IV |
| O | Male | 12;0 | Ataxic | I |
Notes
- a Unilateral spastic cerebral palsy.
- b Bilateral spastic cerebral palsy.
All children had been diagnosed and classified by Himmelmann et al. (2010) based on the dominant symptom according to the Surveillance of Cerebral Palsy in Europe (2000) into unilateral spastic CP (USCP), bilateral spastic CP (BSCP), dyskinetic CP and ataxic CP (table 1). Gross motor function was classified into the five levels of the Gross Motor Function Classification System (GMFCS) (Palisano et al. 1997): level I—walks without limitations; level II—walks with limitations; level III—walks with adaptive equipment assistance; level IV—self-mobility is limited, likely to be transported in a wheelchair; and level V—transported in a wheelchair (table 1).
Materials
Narrative ability
The BST (Renfrew 1997; Svensson and Tuominen-Eriksson 2002), a standardized story-retelling test (for ages 3;9–8;5), consisting of a storybook with 12 cartoon pictures and no written words, was used. Children were told the story orally by the first author while looking at each picture. The children were asked to retell the story as closely to the original as possible using the pictures as prompts. Orthographical transcription was done in accordance with the Swedish manual (Svensson and Tuominen-Eriksson 2000). The BST provides a norm-referenced Information score (BSTinf), i.e. the number of relevant pieces of information given by the child. According to the manual, the Information sub-score reflects the child's ability to comprehend the original story. The five longest sentences, in terms of number of words, were selected in each sample and the mean value was calculated as Sentence Length (BSTsl). As a measure of expressive grammatical complexity, the number of Subordinate Clauses (BSTsc) was calculated and compared with norms. The maximum Information score is 54, and the scores for mean sentence length of the five longest sentences and number of subordinate clauses vary depending on the speech and language capacity of the children performing the test.
The NAP (Bliss et al. 1998) was used for further analysis of the children's narrative ability on six dimensions: Topic Maintenance—how well utterances in a narrative relate to a central topic; Event Sequencing—the presentation of events in chronological or logical order; Explicitness—the sense-making process of discourse coherence, the aspects of the completeness and elaboration of a narrative; Referencing—adequate identification of individuals, features and events; Conjunctive Cohesion—conjunctions that link utterances and events in a narrative; and Fluency, where dysfluencies are defined as lexical or phrasal interruptions in utterances. Finally, a sum score of the different dimensions was calculated, NAP total. The child's achievement on each dimension is assessed and assigned one to three points: 1 point = inappropriate, 2 points = variable, and 3 points = appropriate.
Receptive grammar and vocabulary
Receptive grammar was assessed with the Test for Reception of Grammar Version 2 (TROG-2; Bishop 2003, Sw. version 2009), which identifies difficulties in various aspects of grammatical understanding at sentence level. The test instrument involves matching a total of 80 orally presented sentences with one of four pictures. Normative data are available for Swedish children from ages 4 to 16 (TROG-2; Bishop 2003, Sw. version 2009).
Receptive vocabulary was assessed using the Peabody Picture Vocabulary Test (PPVT-IV; Dunn and Dunn 2007). This is a norm-referenced instrument with 228 test items. For each item, the examiner says one of four possible words on each page and the child responds by selecting the picture that best represents that word. As there are no standardized Swedish norms available, the US norms (2:6–90+ years) were used.
Non-verbal cognitive level
Non-verbal cognitive level was measured using Raven's progressive matrices, coloured version (RCPM) (Raven et al. 1986). The RCPM test consists of 36 items, each containing a picture with a pattern with one part removed. There are six pictured inserts, one of which contains the correct pattern, and the child has to point to the piece considered to be the correct one. The maximum score is 36. This test has been used frequently in studies of children with speech and language impairment (e.g. Larsson and Dahlgren Sandberg 2008). Since there are no standardized norms for the RCPM test on Swedish children, the British norms were used.
Memory
Visual memory span
In accordance with previous research (e.g. Dodwell and Bavin 2008), and since the material used for eliciting narratives was composed of pictures, visuo-spatial memory was hypothesized to be an important skill for performance on the narrative task. Therefore, the Corsi block-tapping task (CB) (Milner 1971) was used to measure visuo-spatial memory. We used a copy of Milner's original test material, which consisted of nine black cubes (3 × 3 × 3 cm) placed randomly on a 25 × 21 cm black board. The cubes were numbered from one to nine, with the number marked on one side of each cube. The numbers were only visible to the experimenter. The experimenter pointed to a number of blocks in a certain order. Subsequently, the participant was asked to point to the same blocks in their order of presentation. The length of the sequences increased gradually, starting with two and continuing until the participant failed two sequences of the same length. All children were presented both the forward and the backward conditions. The longest sequence repeated back correctly was taken as the participants’ CB forward and backward scores. The forward version was used as a measure of visuo-spatial STM. The backward condition was considered to measure visuo-spatial WM.
Auditory memory span
Auditory memory is usually assessed with the Digit Span (DS) from the Wechsler Intelligence Scale for Children—III (Wechsler 1999), which has proven to have a strong relationship with a number of language variables (e.g. Baddeley and Hitch 1974). The forward version was used as a measure of auditory STM. The children were asked to repeat the numbers in the same order as they were presented to them orally. The test began with a sequence of two numbers, then three etc. The children were asked to repeat the numbers in both the forward and the backward conditions. The backward condition was considered to measure auditory WM. The longest sequence repeated correctly by the child was taken as the digit span forward and digit span backward.
Theory of mind (ToM)
The FB items of the two story tests Kiki and the Cat (Lewis 1994) and Birthday Puppy (Sullivan et al. 1994) were used. Both stories were translated into Swedish by the second author. Pictures were presented to the child and a story was told. The pictures involved the understanding of the FB (first and second order) of the central character in the stories. The child was asked five probe questions to ensure that he/she was actively processing the story and had encoded and remembered the key events in the story correctly. Finally, there were two test questions (questions about an individual's concern about other persons’ mental states): the second-order ignorance question and the second-order FB question. The first story, Kiki and the Cat, is a short story about a girl and a cat where the girl laid the cat in the bedroom. While Kiki was in the living room the cat slips out through the window and lies down to sleep in another room. At the end of the story the child is asked to point at the room where the girl would go to find the cat. The second story, Birthday Puppy, is a story about a mother who wants to surprise her son and therefore deliberately misinforms him about his birthday present, but her son discovers the true birthday present without her knowing. Later the grandmother asks the mother if her son knows what he is getting for his birthday (second-order ignorance) and then what the child thinks he is getting (second-order belief). A correct response on the questions yielded 1 point. Hence, a maximum of 2 points could be awarded.
Procedure
All tests and test items were presented in the same order for each child. The assessments were made by the first author and took place at the children's schools, except for one child (Child F) who was assessed at home. For three children with severe gross motor problems (G, H and N), the personal assistants were present at the time of the assessment but did not interact with the children. The narratives were videotaped using a SONY Handicam DCR-TRV50E and an external microphone. The assessment of narrative ability (BST), receptive grammar (TROG-2), receptive vocabulary (PPVT-IV), non-verbal cognitive level (RCPM) assessments of memory functions (Corsi block-tapping task and Digit Span) and the ToM tests (Kiki and the Cat and Birthday Puppy) followed the procedures specified in the respective manuals and articles.
Statistical analysis
SPSS software (version 19.0) was used. Non-parametric methods were used due to sample sizes. Association between variables was evaluated with Spearman's rank correlation coefficient. The level of significance was set at .05.
Earlier research (e.g., Larsson and Dahlgren Sandberg 2008; Holck et al. 2011) on retelling ability in children with CP has guided us to choose the variables in this study. To reduce number of variables, composite scores were computed to create the two variables expressive and receptive language. The individual test scores (age-equivalents) were transformed to z-scores and used in the further analyses. Two composite scores were thus computed: Expressive language (the scores of BST Sentence Length and BST Subordinate Clauses) and Receptive language (the scores of TROG-2 and PPVT-IV). These scores were used together with Non-verbal cognitive level (Raven's progressive matrices), Visual memory (Corsi block forward and backward) and Auditory memory (Digit Span forward and backward) to examine the relationship between these variables and narrative ability operationalized as NAP total score.
Reliability
The first author performed the initial scoring procedure for all three subtests of the BST. Inter-rater reliability was assessed by the first author and another speech–language pathologist (SLP) who had no knowledge of the project. They checked the scoring for six (40%) of the BST transcriptions, and a correlation analysis was performed. The inter-rater reliabilities for the BST subtests were significant: Information (r = .940) p < 0.1), Sentence Length (r = .829 p < .05) and Subordinate Clauses (r = .904 p < .05). Inter-rater reliability assessments were conducted also for the NAP analysis. For the NAP analysis, the first author and an independent coder (another SLP) jointly coded transcripts for eight children with speech and language impairment (children not included in this study) to practise the coding scheme. Disagreements were resolved through discussion. Following this, all transcripts were coded independently by the first author and the independent coder to determine inter-rater reliability. The correlation analysis showed correlations varying from .46 to 1.0, with the lowest value for Event Sequencing and Fluency and the highest for Topic Maintenance.
Ethics
The research ethics committee in Gothenburg approved the study and the parents provided written informed consent.
Results
Narrative ability
The Bus Story Test (BST)
For the BST measure Information mean raw score was 20.6 (SD = 11.7; range = 4–42), corresponding to an age-equivalent of 5;2 years (range = 3;8–8;5). Mean raw score for Sentence Length was 8.6 words (SD = 3.08; range = 4–14), corresponding to an age-equivalent of 6;11 years (range = 3;8–8;5). Mean raw score for Subordinate Clauses was 3.3 (SD = 3.11; range = 0–9), corresponding to an age-equivalent of 5;9 years (range = 3;10–8;5). See table 2 for the children's individual BST results.
| Chronological | Bus Story Testa | Corsi block | Corsi block | Digit span | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| age | forward | backward | forward | Digit span | ||||||||
| Child | Gender | (years;months) | Inform | sent length | sub-clause | TROG-2a | PPVT-IVa | Ravena | STMb | WMb | STMb | backward WMb |
| A | Female | 9;6 | 3:9 | 3:9 | 3:9 | 4:1 | 6:1 | 5 | 4 | 2 | 3 | 3 |
| B | Male | 9;9 | 5:3 | 5:7 | 4 | 6:6 | 8:7 | 9 | 4 | 4 | 4 | 3 |
| Cc | Male | 12;8 | 8:5 | 6:10 | 5:2 | 9:8 | 11:2 | 8 | 5 | 2 | 6 | 2 |
| D | Male | 11;11 | 3:9 | 4:6 | 4:6 | 5:10 | 7:5 | 9 | 4 | 3 | 4 | 2 |
| E | Female | 12;9 | 8:5 | 8:5 | 5:6 | 9:8 | 13:1 | 10:6 | 6 | 4 | 5 | 3 |
| F | Female | 12;8 | 6:1 | 6:10 | 5:6 | 5:10 | 7:2 | 11:6 | 6 | 4 | 4 | 3 |
| G | Female | 9;8 | 7:7 | 7:4 | 4 | 14:7 | 11:1 | 6 | 2 | 3 | 5 | 3 |
| H | Male | 10;7 | 3:9 | 3:9 | 3:9 | 4:0 | 4:7 | Visual | Visual | Visual | 3 | 0 |
| impairment | impairment | impairment | 3 | 0 | ||||||||
| I | Male | 11;3 | 3:9 | 4:6 | 3:9 | 6:0 | 6:10 | 9 | 2 | 0 | 5 | 2 |
| Jc | Male | 11;8 | 4:5 | 3:9 | 3:9 | 5:10 | 5:4 | 6 | 3 | 3 | 3 | 3 |
| K | Male | 11;10 | 5:10 | 7:10 | 5:8 | 9:8 | 16:11 | 11 | 5 | 4 | 5 | 2 |
| L | Female | 9;2 | 4:7 | 6:4 | 3:9 | 6:11 | 6:9 | 6 | 5 | 3 | 6 | 3 |
| M | Female | 10;1 | 4:7 | 7:4 | 4:4 | 7:5 | 6:4 | 7 | 3 | 2 | 4 | 2 |
| N | Male | 9;2 | 4:10 | 5:1 | 4:3 | 4:0 | 5:10 | –d | 3 | 2 | 4 | 0 |
| O | Male | 12;1 | 3:9 | 6:4 | 3:9 | 5:10 | 7:8 | 8 | 5 | 3 | 8 | 5 |
Notes
- a Age-equivalents.
- b Raw scores.
- c Bilingual.
- d Below available norms (5.6–11.6 years).
The Narrative Assessment Profile (NAP)
- Topic Maintenance: Thirteen of the children displayed appropriate topic maintenance. Two children added off-topic contributions.
- Event Sequencing: All children but two reached the maximum score.
- Explicitness: Nine of the 15 children were unable to make a fully explicit and elaborated narrative.
- Referencing: Nine of the 15 children had problems with referencing.
- Conjunctive Cohesion: Six children demonstrated appropriate use of conjunctions, the use varied for four children and five children showed inaccurate use of conjunctions.
- Fluency: Twelve out of 15 children had an inappropriate or variable ability to create narratives without dysfluencies.
| Child | Topic | Event | Explic | Refere | Conjunctive | NAP | |
|---|---|---|---|---|---|---|---|
| (N = 15) | maintenancea | sequencinga | itnessa | ncinga | cohesiona | Fluencya | totala |
| A | 3 | 2 | 1 | 1 | 1 | 3 | 11 |
| B | 3 | 3 | 1 | 1 | 1 | 1 | 10 |
| C | 3 | 3 | 3 | 3 | 3 | 2 | 17 |
| D | 3 | 3 | 1 | 1 | 1 | 1 | 10 |
| E | 3 | 3 | 3 | 3 | 3 | 3 | 18 |
| F | 3 | 3 | 1 | 1 | 2 | 1 | 11 |
| G | 3 | 3 | 2 | 2 | 3 | 3 | 16 |
| H | 3 | 3 | 1 | 1 | 1 | 1 | 10 |
| I | 2 | 2 | 1 | 1 | 2 | 1 | 9 |
| J | 2 | 3 | 2 | 1 | 1 | 1 | 10 |
| K | 3 | 3 | 3 | 2 | 3 | 3 | 17 |
| L | 3 | 3 | 2 | 3 | 3 | 2 | 16 |
| M | 3 | 3 | 1 | 2 | 3 | 2 | 14 |
| N | 3 | 3 | 1 | 1 | 2 | 1 | 11 |
| O | 3 | 3 | 1 | 2 | 2 | 2 | 13 |
Note
- a Children's achievement on each dimension is one to three points: 1 point = inappropriate, 2 points = variable, and 3 points = appropriate.
Receptive grammar and vocabulary
TROG-2
Mean raw score was 59 out of a possible 80 (SD = 19.26; range = 18–76; median = 66), corresponding to an age-equivalent of 7;1 years (range = 3;11–14;7) (chronological mean age = 11;0 years). A majority of the children had problems with receptive grammar according to TROG-2. See table 2 for individual TROG-2 results.
PPVT-IV
Mean raw score was 125 out of a possible 228 (SD = 33.84; range = 73–193; median = 118), corresponding to an age-equivalent of 8;4 years (range = 4;7–16;11). Also here, a majority of the children had problems with receptive vocabulary according to PPVT-IV. See table 2 for individual PPVT-IV results.
Non-verbal cognitive level
Mean raw score was 20 out of a possible 36 (SD = 8.37; range = 6–33; median = 21), corresponding to an age-equivalent of 8;2 years (range = 5;0–11;6). See table 2 for individual (RCPM) results.
Memory
For visual memory span forward, the mean raw score was 4.1 (SD = 1.33; range = 2–6), while the span for WM was 2.8 (SD = 1.12; range = 0–4). For auditory memory span forward, the mean raw score 4.6 (SD = 1.3; range = 3–8), while the digit span WM was 2.4 (SD = 1.24; range = 0–5). See table 2 for individual memory results.
Theory of mind (ToM)
All children in the study except one (Child N) passed the first-order false belief ToM test Kiki and the Cat. Four children did not pass the probe questions (H, I, J and N) in the second-order ToM test Birthday Puppy and were therefore excluded. All remaining 11 of the 15 children passed the first second-order ignorance question. For the second-order FB question, out of a maximum of 2 points, the mean score was 1.36 (SD = .51).
Correlations
According to the Spearman rank correlation test, there were significant relationships between the measures of TROG-2, PPVT-IV and the second-order ToM test, on the one hand, and the three BST measures, on the other. Digit span forward and Raven's coloured matrices correlated only with the BST sub-score of Sentence Length (table 4). There was a significant relationship between the explicitness dimension of the NAP and the BST Information score (r = .532*; n = 15; p = .04). No other dimension of the NAP correlated with any of the BST scores.
| 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 BSTinf | 1 | |||||||||||
| N | 15 | |||||||||||
| 2 BSTsl | .780** | 1 | ||||||||||
| N | 15 | 15 | ||||||||||
| 3 BSTsc | .806** | .912** | 1 | |||||||||
| N | 15 | 15 | 15 | |||||||||
| 4 TROG-2 | .887** | .907** | .849** | 1 | ||||||||
| N | 15 | 15 | 15 | 15 | ||||||||
| 5 PPVT-IV | .701** | .790** | .716** | .845** | 1 | |||||||
| N | 15 | 15 | 15 | 15 | 15 | |||||||
| 6 CB, forward | .418 | .441 | .446 | .349 | .463 | 1 | ||||||
| N | 14 | 14 | 14 | 14 | 14 | 14 | ||||||
| 7 CB, backward | .498 | .446 | .372 | .467 | .525 | .600* | 1 | |||||
| N | 14 | 14 | 14 | 14 | 14 | 14 | 14 | |||||
| 8 DS, forward | .301 | .615* | .340 | .576* | .666** | .338 | .041 | 1 | ||||
| N | 15 | 15 | 15 | 15 | 15 | 14 | 14 | 15 | ||||
| 9 DS, backward | .243 | .123 | .259 | .230 | .261 | .332 | .464 | .258 | 1 | |||
| N | 15 | 15 | 15 | 15 | 15 | 14 | 14 | 15 | 15 | |||
| 10 ToM second belief | –.719* | –.717* | –.667* | –.784** | –.777** | –.402 | –.570 | .402 | .329 | 1 | ||
| N | 11 | 11 | 11 | 11 | 11 | 11 | 11 | 11 | 11 | 11 | ||
| 11 RCPM | .487 | .557* | .475 | .577* | .709** | .562* | .602* | .345 | .230 | –.570 | 1 | |
| N | 15 | 15 | 15 | 15 | 15 | 14 | 14 | 15 | 15 | 11 | 15 |
Note
- *p < .05; **p < .01; ***p < .001
| 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | |
|---|---|---|---|---|---|---|---|---|
| 1. NAP total | 1 | |||||||
| N | 15 | |||||||
| 2. Expressive Language | .825** | 1 | ||||||
| N | 15 | 15 | ||||||
| 3. Receptive Language | .719** | .839** | 1 | |||||
| N | 15 | 15 | 15 | |||||
| 4. Visual Memory, forward | .511 | .507 | .294 | 1 | ||||
| N | 14 | 14 | 14 | 14 | ||||
| 5. Visual Memory, backward | .254 | .502 | .486 | .600* | 1 | |||
| N | 14 | 14 | 14 | 14 | 14 | |||
| 6. Auditory Memory, forward | .588* | .624* | .623* | .338 | .041 | 1 | ||
| N | 15 | 15 | 15 | 14 | 14 | 15 | ||
| 7. Auditory Memory, backward | .184 | .234 | .267 | .332 | .464 | .258 | 1 | |
| N | 15 | 15 | 15 | 14 | 14 | 15 | 15 | |
| 8. Raven's matrices | .058 | .490 | .295 | .507 | .576* | .128 | –.256 | 1 |
| N | 13 | 13 | 13 | 13 | 13 | 13 | 13 | 13 |
Narrative ability was defined as NAP total. Correlations were computed to examine the relationship between narrative ability and possible background factors. There were statistically significant relationships between the NAP total score, on the one hand, and the composite scores expressive language (r = .825**; n = 15; p = .01) and receptive language (r = .719**; n = 15; p = .03) and with auditory memory, forward condition (r = .588**, n = 15, p = .021), on the other.
Discussion
The children with CP and speech impairment in this study demonstrated low narrative ability as indicated by the low scores on all three BST narrative subtests. The BST information scores were lower than the BST results for Sentence Length and Subordinate Clauses. A possible explanation for the lower BST information results was that the children had to retell the bus story in chronological order to score, i.e. give the right information for each of the 12 pictures in order, and this was too demanding for 11 of the 15 children in the study. Children with CP and speech impairment may have a range of difficulties such as varying degrees of motor problems and cognitive impairments (e.g. Nordberg et al. 2014). In the present study, this was evident in the wide range of scores on the language ability tests, both across and within the children. The retelling ability score for BST information ranged from an age-equivalent of 3;9–8;5 years. Child G (9;8 years) performed above average in understanding of grammatical structures (TROG-2) with an age-equivalent of almost 15 years. Child K (11;10 years) scored far above average for receptive vocabulary (age-equivalent 16;11 years). On the other extreme, Child N (9;2 years) performed as a 4-year-old on the TROG-2 test.
The speech production/articulation of the children in this study has been investigated and described in detail in a previous study. More than half of the children had large problems with the articulation of consonants, and the majority of them were rated as having dysarthria (Nordberg et al. 2014). It is important to take the children's speech impairments and low non-verbal cognitive level into account when interpreting the low scores on the narration task. Speech impairment was an inclusion criterion in the present study and the mean mental age was 7;11. The children in the earlier referred study by Holck et al. (2011) were indeed speaking, but had varying degrees of language impairment (mainly affecting the phonological ability) and average cognitive level. They performed just a little inferior to TD children on both linguistically and cognitively related narrative measures. The dissimilar retelling ability scores in this present study and the Holck et al. (2011) study could thus be expected due to the differences between the participants’ non-verbal cognitive level and, to some extent, their various speech and language abilities. On the contrary, the two studies showed similar results of the NAP analysis. The children had problems with the NAP dimension Explicitness, i.e. difficulties creating narratives containing sufficient information for the listener. Despite their differing ability to speak and level of non-verbal cognitive ability, they performed equally poorly on this NAP dimension. Yet, Explicitness requires an ability to understand others’ minds and in both studies the children showed limited ToM ability, which could explain the low results.
Memory is one of the cognitive abilities commonly regarded as crucial to story recall (e.g. Dodwell and Bavin 2008). The children in this study had difficulties in all memory functions, both visual and auditory, and scored low on all story retelling subtests. The low non-verbal cognitive profile of the children is probably an important factor contributing to their low memory results. The children performed best on auditory STM and had the greatest difficulty with auditory WM. Accordingly, there were problems with the auditory memory storage component, which is an important factor in the retelling task since it demands keeping material in memory for future use while retelling the story. A sign of this is their difficulties with the NAP dimension Fluency, possibly caused by their reduced ability to hold information in mind. The children with CP in the present study used speech as a means of communication. For some children, especially for those with severe speech impairment, use of AAC may be a better way to communicate. It is important to note that using AAC may lead to other difficulties with the memory functions than for speaking children. For example, an external aided AAC device, containing a lot of symbols at the same time, places unique demands on the children's memory capacity (Thistle and Wilkinson 2013).
Half of the school-aged children in the present study had marked problems with story retelling ability and receptive grammar, children who scored low on BST also scored low on TROG-2. Accordingly, covariance was found between the production aspects and the comprehension aspects of language among children with CP in this study. A number of researchers have stressed the importance of examining comprehension in children who have problems with narratives (e.g. Reuterskiöld Wagner et al. 1999, Leinonen et al. 2000, Norbury and Bishop 2003) since they have demonstrated a covariance between comprehension and the ability to produce a narrative, shown in the present study as well.
The performance of the second-order ToM task correlated significantly with retelling ability in this study. ToM has also been studied by Dahlgren et al. (2010) in 16 children with CP without speech (mental age = 4–9.5 years) and they had worse performance on ToM tasks than could be expected from TD children. The low ToM results were discussed in terms of poor language abilities. This has also been found for children with other language problems, such as specific language impairment (SLI) (Botting and Conti-Ramsden 1999). In the study by Holck et al. (2010), with 10 speaking children with CP (the same children as in the study Holck et al. 2011, on narrative ability), they performed worse than matched TD children on ToM. The ability to predict and explain other people's actions (ToM) has been proven to be difficult, for the children with CP without speech (Dahlgren et al. 2010), for the speaking children with CP (Holck et al. 2010), and for the speaking children with CP and speech impairment in this present study. ToM affects interaction and communication and is important for retelling something in a way that is understandable to others. This issue needs to be further investigated in larger groups of children with CP.
Today, assessment of narrative ability is not regularly used by Swedish speech and language pathologists because they are too time-consuming. However, it is clear that oral narratives together with assessment of receptive grammar and vocabulary capture subtle language difficulties and could be a good predictor of long-term language skills (Bishop and Edmundsson 1987; Stothard et al. 1998; Miniscalo et al. 2007). The BST test will not capture all problems the children with CP and speech impairment may have. Preferably it can be used as a screening instrument and if the scores turn out to be low on BST information complementary assessment of core language abilities should be made. Low BST scores of sentence length may require further assessment of memory functions and performance with few subordinate clauses will focus the attention on problems with syntactic complexity.
The result of the NAP analysis (Bliss et al. 1998) gives a multi dimensional profile of the speaker's nature of narrative discourse. The significant relationships between the NAP total result and the composite scores of both expressive and receptive language in this study illustrate the importance of production and comprehension of language when retelling a story. Also, the two core language composites were associated with auditory memory, which is in line with earlier research (e.g. Baddeley and Hitch 1974). As these results indicate assessment of core language abilities and memory functions (especially auditory memory), is important for this group of children. ToM and non-verbal cognitive level are important cognitive aspects to consider, as well.
The sample size in this study was small, due to the fact that the group of children with CP and speech impairment is very limited when selected, as in this study, from a population based age-grouped cohort. This calls for caution in interpreting the results. The lack of a control group is a limitation; instead norm-referenced tests are used, even if the norms are for children without disabilities. In addition, since the sample is small, extremely low scores can have influenced the results. At the same time there are big individual variations in the data, which may partly compensate for this.
The children with CP and speech impairment in this study demonstrated a high rate of narrative problems. A greater understanding of language and cognitive ability in children with CP and speech impairment is needed in order to find appropriate support for these children to communicate and be able to participate in society on equal terms as their typically developed peers.
Acknowledgments
This investigation was supported by grants from Habilitation & Health, Region Västra Götaland, Sweden, the Petter Silfverskiöld Foundation, the Queen Silvia Jubilee Foundation and the Stinger Foundation. Our special thanks go to the children and their families and to Emilia Carlsson for the reliability assessment.
Declaration of interest
The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.
