A longitudinal case study of six children with autism and specified language and non-verbal profiles

ABSTRACT

Language skills as well as general cognitive skills show a considerable variation in children with autism spectrum disorder (ASD). In previous studies, at least three profiles based on these skills have been suggested; autism with language and non-verbal cognitive skills within the average/normal range (ALN), autism with language disorder (ALD) without concurrent non-verbal cognitive disability, and autism with language disorder and cognitive disability, i.e. autism with a more general delay (AGD). The aim of the present longitudinal case study is to illustrate these three groups more thoroughly by presenting the developmental trajectories of children belonging to each profile. Six children were chosen based on their language and cognitive profiles from the first age 3-year assessment. They came from a larger group of children with ASD identified by autism screening at child health-care centres at age 2.5 years. These six children represent one boy and one girl from each of the three subgroups ALN, ALD and AGD, and were assessed a second time at age 5 and a third time at age 8 years, regarding expressive and receptive language skills, autistic severity and non-verbal cognitive skills. Although preliminary, our results indicate a rather stable developmental trajectory from age 3 to 8 years characterising children with autism based on language and non-verbal cognitive functioning. Thus, in order to help intervention planning and increase predictions of outcome, it seems important to specify both linguistic and cognitive level already at the first assessment in children with ASD.

KEYWORDS:

• Language skills
• non-verbal cognitive skills
• ASD
• development

Autism spectrum disorder (ASD) is a congenital or early-acquired complex neurodevelopmental disorder characterised by difficulties in social interaction, impaired social communication and restricted behavioural patterns (Association, 2013; Wing & Gould, 1979). The symptoms must be present in the early developmental period, typically before age of three, and usually persist throughout life (APA, 2013). In the general population the prevalence of ASD is about 0.8% to 2% (Baird, Simonoff, Pickles et al., 2006; Coleman & Gillberg, 2012; Nygren et al., 2012).

Early signs of ASD often appear in the first year of life, making the toddler years a crucial period to identify symptoms of ASD. Concern about delayed speech and language development is common and often the reason why parents seek referral and clinical examinations of their child in the first place (Dahlgren & Gillberg, 1989; De Giacomo & Fombonne, 1998). Thus, children with ASD are often identified due to their delay in language milestones. Many children lack babbling or use very little babbling during the first year of life (Luyster et al., 2011; Oller et al., 1999; Paul et al., 2011). In addition, many children with ASD show delays and a slower development rate in speech acquisition compared with peers without ASD, saying their first words at 24 months and producing their first sentences at 48 months (Alin-kerman & Norberg, 1991; Gernsbacher et al., 2016; Howlin, 2003; Tager-Flusberg et al., 2005). Moreover, some toddlers (10–25%) with ASD may suddenly lose the language skills they earlier mastered (Barger et al., 2013; Fernell et al., 2010; Kantzer et al., 2013; Thompson et al., 2019). This sudden language regression usually appears during the second year of life, with reports of both losses in productive language use and more subtle decreases in vocalisations and social communication (Lord et al., 2004; Ozonoff et al., 2010). Regression in language skills seems to be a unique marker for ASD (Lord et al., 2004) and affects approx. 20% of these children (Thompson et al., 2019).

Difficulties with social communication or pragmatic ability, i.e., the conventions and principles regarding how language is used in context, is one of the main symptoms of autism (APA, 2013). On the other hand, difficulties in understanding and producing language both in terms of receptive and expressive language function vary greatly between individuals with autism (Boucher, 2012; Kjelgaard & Tager-Flusberg, 2001).

Several studies have reported that impaired receptive language is common in children with ASD (Charman, Drew et al., 2003; Kjellmer et al., 2018; Kover et al., 2013; Loucas et al., 2008; Norrelgen et al., 2015). Some studies have shown that receptive language is more impaired than expressive language (Charman, Baron-Cohen et al., 2003; Hudry et al., 2010; Weismer et al., 2010), whereas others have found the opposite pattern (Luyster et al., 2008). Moreover, it has previously been shown that toddlers with ASD have smaller expressive vocabulary than peers (Charman, Drew et al., 2003; Miniscalco et al., 2012) and that some of them have language difficulties similar to those in developmental language disorder (DLD) (Kjellmer et al., 2018; Tager-Flusberg et al., 2003).

Although studies have examined both receptive and expressive language ability, few studies have focused on the relationship between them. A meta-analysis by Kwok et al. (2015) showed that younger children performed roughly −2 SD below on both receptive and expressive language skills compared to children without ASD, while the older children with ASD performed within normal limits on both receptive and expressive language measures. This could suggest that children with ASD tend to catch up over time, or that the study groups tend to be more heterogeneous among younger children compared to the older children with ASD (Kwok et al., 2015).

According to the changes in DSM-5, cognitive and linguistic abilities need to be specified, and dual diagnosis of ASD and “language disorder” (LD) (defined primarily as deficits in vocabulary and grammar) must be assigned when the child meets the criteria for both conditions (ADHD)(APA, 2013).

In summary, both language and non-verbal cognitive skills are important to take into account when diagnosing children with ASD in order to help intervention planning and increase predictions of outcome. Knowledge of the developmental trajectory from childhood into adolescent would be valuable, since children with ASD and LD might need more and specific intervention to deal with their language difficulties (Tager-Flusberg, 2015).

The aim of the present study is to describe language skills and non-verbal skills in children with ASD, and to define whether they also have accompanying LD or non-verbal cognitive disorder or not. By describing these children more thorough at several time points and providing some context of what each case demonstrates we were also interested in how stable these subgroups are over time. These three groups are children with autism with language and non-verbal cognitive skills within the average/normal range (ALN), autism with language disorder (ALD) without concurrent non-verbal cognitive disability, and autism with language disorder and cognitive disability, i.e. autism with a more general delay (AGD).

Method

Participants

The six children (three girls and three boys) participating in the present case study were recruited from a large longitudinal population-based study, performed at the Child Neuropsychiatric Clinic (CNC) in Gothenburg (Kantzer et al., 2013, 2018). The children were initially identified by a general ASD screening at age 2.5 years at the public child health-care centres (CHCs) in Gothenburg, i.e. AUtism Detection and Intervention in Early life, or the AUDIE project (Nygren, Cederlund et al., 2012; Nygren, Sandberg et al., 2012) and were assessed three times during a five year time period. From this original sample of 129 screen positive children, 107 participated at T1 (at age 3 y), 96 at T2 (at age 5 y) and 85 at T3 (at age 8y). Out of this large T3 sample, six children were selected based on their language and cognitive profiles at study intake T1, see Table 1.

Table 1. Participant characteristics from T1 (3 years of age).

	ALN		ALD		AGD
Linguistic and cognitive profile	“Paul”	“Linda”	“Michael”	“Tina”	“William”	“Anna”
Gender	♂	♀	♂	♀	♂	♀
Age (months)	38	29	37	37	39	45
Non-verbal cognitive ability DQ (Griffith’s)	105	89	76	113	86	77
Autism severity^a (ADOS)	4	8	8	2	6	6
Receptive language (z-score) (RDLS III) ^b	50 (1.2)	35 (0.7)	10 (- 5.2)	15 (−5.9)	0 (−4.0)	0 (−4.0)
Expressive language skills (Paris)^c	5	5	2	1	2	4
Vocabulary (Parental rating, SECDI)	381	283	4	12	24	207
Gestures (Parental rating, SECDI)	63	53	35	50	27	37
MLU, rated by parents (Parental rating, SECDI)	8,5	3,7	n.p.	8	1	2,7

ALN = autism and language normal, ALD = autism and language disorder, AGD = Autism and intellectual disorder

^aStanine. ^bRaw scores. ^c1 = no words at all; 2 = a few single words; 3 = a few communicative sentences; 4 = talks a great deal, mostly echolalia, or 5 = talks a great deal, mostly in a communicative fashion. ^d)ADOS severity scores 1–10, 10 = most severe. N.p. = the particular tests were not fulfilled/performed. All included participants were given a figurative name.

Procedure

Two speech and language pathologists (SLPs) at the CNC, assessed all children during at least two occasions at three time points, at T1 (at age 3 y), T2 (at age 5 y) and T3 (at age 8y) during a five-year time period. Each visit lasted approximately 60 minutes. For time point 1 and 2, the children were multi-professionally assessed by an SLP, a neuropsychologist, a special education teacher, and a doctor with several years of training in child neuropsychiatry. At T3, the focus was on linguistic ability/language functioning and the children were assessed by the SLPs with a comprehensive test battery focusing on expressive and receptive language, non-verbal skills, reading and narrative skills. In addition, the parents answered questions about developmental functioning and autistic symptomatology.

Materials

An overview of the materials used at the three time points is presented in appendix 1.

Receptive language ability

Receptive language (language comprehension) was assessed with the Reynell Developmental Language Scales III (RDLS) (Edwards et al., 1997) using Swedish norms (Eriksson & Grundström, 2000; Lindström & Åström, 2000). The results are presented in raw scores and transformed into z-scores (M = 0, SD = 1). The reliability coefficient Kuder-Richardsson is reported to be 0.97 in the manual from U.K.

The receptive language was assessed both at T2 and T3, with the Test for Reception of Grammar-2 (TROG-2) (Bishop, 2009). The test leader presents four pictures to the child, then the child is supposed to match orally presented sentences to the correct picture. The results are presented both in terms of raw scores (number of correctly solved blocks out of a maximum of 20) and standard scores (M = 100, SD = 15) based on Swedish norms. The Cronbach’s alpha is reported to be 0.89 in the Swedish manual.

Vocabulary was assessed at T2 and T3 using the Peabody Picture Vocabulary Test, Third Edition, (PPVT-III) (Dunn & Dunn, 1997). The child listens to a word uttered by the SLP and then points at one of four pictures, the one that best describes the word’s meaning. The test is not standardised for Swedish children in the current ages, and therefore the original U.S. norms were used. The Cronbach’s alpha is reported to be 0.95 in the U.S manual.

Expressive language ability

The expressive language level of each child was rated by the SLP on a scale from 1-to-5 using the PARIS scale (Philippe et al., 1999). The scale range from 1 to 5; where 1 = no words at all; 2 = a few single words; 3 = a few communicative sentences; 4 = talks a great deal, mostly echolalia, or 5 = talks a great deal, mostly in a communicative fashion.

As a measure for expressive grammar at T3, the subtest “recalling sentences” from the Clinical Evaluation of Language Fundamentals – 4 (CELF-4) (Semel et al., 2013) was used (Klem et al., 2015). “Recalling sentences” consists of 24 sentences. The participant has to repeat each sentence produced by the test leader, resulting in a score from 0 (>4 errors) to 3, (no errors), max score of 72. The results are presented in raw scores and scaled scores (around a normative M = 10, SD = 3; Swedish norms). The Cronbach’s alpha is reported to be 0.89 in the Swedish manual.

Phonology and non-word repetition

Consonant inventory, the number of different consonants the child produced out of 18 possible consonants was calculated, based on either of the expressive language materials. The SLP used broad phonetic transcription according to the International Phonetic Alphabet (International, Phonetic Association, 1999). Non-word repetition was assessed with 30, one to five syllable non-words that conform to Swedish phonotactics (Radeborg et al., 2006). The children repeated the non-words after the SLP’s oral presentation. Each repeated non-word was rated as correct or incorrect. Norms are available for children aged 4–6 years. Radeborg et al. report a Cronbach’s alpha of 0.74.

Narrative skills

Oral story retelling was assessed using The Bus Story Test (BST)(Renfrew, 1997; Svensson & Tuominen-Eriksson, 2000) at both the T2 and the T3 follow-up. The BST consists of a 12-cartoon colour pictures storybook about a “naughty” bus. The test leader reads the story and then the child is asked to retell the story while looking at the pictures. All stories were audio-recorded and orthographically transcribed according to the Swedish manual. The test is standardised for Swedish children within the age range of 3.9–8.5 years.

Pragmatic/communicative functioning

The children’s communicative functioning was rated by their parents using the Children’s Communication Checklist – second version (Bishop, 2003). The questionnaire consists of 70 items, in total eight domains. Four domains representing speech, grammar, vocabulary and discourse (speech, syntax, semantics, coherence) and four domains concern communication (initiation, stereotyped language, use of context and non-verbal communicative ability). Each domain corresponds to a scaled score with a normative mean of 10 and an SD of 3.

Assessment of cognitive/developmental level

A clinical child neuropsychologist assessed children’s general cognitive and developmental level using the Griffiths’ developmental scales (GDS) (Alin-kerman & Norberg, 1991). The test includes six subscales; the total score (M = 100, SD = 15) from the subscales provides a developmental quotient (DQ) which is used here. McLean et al. (1991) report adequate psychometric properties, both in terms of internal consistency reliability (Cronbach’s alpha >0.96) and construct validity according to correlation patterns with other tests of cognitive functioning (McLean et al., 1991). The non-verbal skills, the performance intelligence quotient (PIQ), at T2 was calculated from the assessment with WPPSI-III (Wechsler, 1991).

At T3, follow-up the subtest Matrix reasoning from the Wechsler abbreviated scales of intelligence (Wechsler, 1999) was used assessing non-verbal cognitive skills. The results are presented as raw and T-scores (M = 50, SD = 10) based on American norms, no Swedish norms are available.

Autism severity and symptomatology

The Autism Diagnostic Observation Schedule – Generic (ADOS) (Lord et al., 2000) is a standardised, semi-structured play-based assessment of communication, reciprocal social interaction, play, and behaviour. Either module 1 or 2 was used, based on the expressive language level of the child, and from these data calibrated severity scores were calculated (scores from 1 to 10) (Gotham et al., 2009; Hus et al., 2014). Higher scores indicate increased autistic symptom severity. The cut-off for ASD is 4, and 6 is for autistic syndrome.

At T3, all the parents completed the Autism Spectrum Screening Questionnaire (ASSQ) (Ehlers et al., 1999) to assess autistic symptomatology. The questionnaire contains 27 items/statements with a three-level Likert scale. The test–retest reliability is reported to be very high for parent reports (r = .96). Validity was established by Ehlers et al. (1999) and by Posserud et al. (2009) by showing a clear correspondence between the total ASSQ score and a clinical diagnosis of ASD. A cut-off for ASD of >18 has been suggested (Ehlers et al., 1999; Posserud et al., 2009).

Adaptive communication and social functioning

The Communication and Socialization domains of the Vineland Adaptive Behaviour Scales (VABS) (Sparrow et al., 2005) was administered by a child neuropsychologist in a face-to-face interview with one or both parents. Standard scores have a normative value of M = 100, SD = 15.

Group assignment

The six children were divided into three subgroups according to DSM-5 specifiers of language ability and non-verbal ability based on their test results from T1; (1) the ‘autism language normal’ children (ALN) obtained z-scores on the RDLS receptive (Edwards et al., 1997) of ≥-1.3 SD and non-verbal ability standard scores of ≥80, (2) the ‘autism language disorder’ children (ALD) obtained RDLS receptive z-scores of <-1.3 and non-verbal ability standard scores of ≥80, and (3) the ‘autism and general delay’ children (AGD) with z-scores on the RDLS receptive of <-1.3, and non-verbal ability scores of <80 equals the 8^th percentile. These cut-offs were chosen based on the study of (Norbury et al., 2016) (Norbury et al., 2016) of the incidence of language disorder among 5-years-olds (i.e. 8%) and from Loucas et al. who defined ALD in individuals with non-verbal ability of ≥80 (Loucas et al., 2008). In this study, we have used figurative names, to ensure anonymity. See Table 1 for participant’s characteristics at T1.

Ethics

The study received ethical approval from the Regional Ethical Review Board in Gothenburg, Sweden (case number 723–13). All the parents of the participating children provided oral and written informed consent.

Results

Each case is described in detail below and represents true cases taken from Kantzer et al. (2013), Kantzer et al. (2018)). The results of the six children’s expressive and receptive language measures at the three time points are presented and illustrated in Figures 1 and Figures 2.

Figure 1. Receptive language for all six children at T1 (3 years of age) RDLS z-score, T2 (5 years of age) TROG-2 z-score and T3 (7–8 years of age) TROG-2 z-score.

Figure 2. Expressive language for all the six children at T2 (5 years of age) and T3 (7–8 years of age) measured with Bus Story Test, the subscore Sentence Length. Note that there is one missing data point for Paul (T2) and one for William (T3).

“Paul” Case 1 (ALN)

“Paul”, was 3.1 years old at his first clinical assessment, see Table 2. He had a family history of ASD and other psychiatric disorders and received an ASD diagnosis at this first assessment. Paul spoke in sentences (PARIS level 5), scored high on RDLS receptive (Edwards et al., 1997) (50 raw score). Two years later, at T2, when he was 5.7 years old he was diagnosed with Asperger Syndrome i.e. (I. C. Gillberg & Gillberg, 1989). His Performance IQ (PIQ) and language test results were average or above average. He had an age-appropriate speech with 18 established consonants out of 18 in his repertoire and no phonological simplification processes, see Table 2. Then, at the T3 school-age follow-up when he was 8.3 years old, he went to first year of primary school (Table 3). He was very interested in reading and read one book a day according to his parents. The SLP assessment revealed that his language skills still were very good; receptive grammar was average (standard score 111 on the TROG-2) but receptive vocabulary (standard score 150 on the PPVT-III) and his ability to recall sentences (scaled scores 19 on Recalling sentences on the CELF-4) were considerably above average. For the Bus story test, he scored within the expected age norms on all three measures. His non-verbal skills measured with the matrices (T-score 47) showed an average result. However, despite his good language functioning his parents rated his autism symptoms higher (ASSQ = 18) (Ehlers et al., 1999) indicating that his autism presentation had become more obvious, and his peers called him “the little professor”.

Table 2. Data from T2 (5 years of age).

	ALN		ALD		AGD
Linguistic and cognitive profile	“Paul”	“Linda”	“Michael”	“Tina”	“William”	“Anna”
Age (years)	5.7	4.6	4.8	5.4	4.9	5.0
Non-verbal cognitive ability (PIQ)	125	129	91	100	75	76
Receptive language (standard scores TROG-2)	119	113	68	55	55	71
Receptive vocabulary (standard scores PPVT III)	135	117	84	64	46	90
Expressive language skills (Paris) ^b	5	5	4	3	3	5
Consonant inventory	18	17	14	17	10	17
Autism Severity (ADOS sev tot)	5	3	9	6	9	8
VABS total	74	84	82	80	65	74
VABS socialization	66	85	85	74	59	77
VABS communication	95	93	85	83	59	87
Information (BST)	n.p.	22	3.4	2	1.6	9
Sentence Length (BST)	n.p.	5.0	3.4	2.6	1.6	3.8
Subordinate Clauses (BST)	n.p.	0	0	0	0	0

^a1 = no words at all; 2 = a few single words; 3 = a few communicative sentences; 4 = talks a great deal, mostly echolalia, or 5 = talks a great deal, mostly in a communicative fashion. N.p. = the particular tests were not fulfilled/performed

Table 3. Data from T3 (7-8 years of age).

	ALN		ALD		AGD
Linguistic and cognitive profile	“Paul”	“Linda”	“Michael”	“Tina”	“William”	“Anna”
Age (years)	8.3	6.9	7.8	7.4	7.3	7.3
Non-verbal ability (Matrices T-score)	47	66	63	53	28	41
Receptive language (standard scoresTROG-2)	111	110	88	55	55	55
Receptive vocabulary (Standard scores PPVT III)	150	125	77	47	40	75
Expressive/Recalling sentences (scaled scores CELF-4)	19	15	5	4	1	7
Autism Symptomatology (ASSQ)	18	32	10	20	20	31
Non-word repetition (max 30)	30	23	23	17	Np	18
Information (BST)	45	26	18	6	n.p.	15
Sentence Length (BST)	9.4	8.2	8.6	5.2	n.p.	5.6
Subordinate Clauses (BST)	3	2	3	0	n.p.	1

“Michael” Case 2 (ALD)

“Michael” was 2.3 years old at his first assessment. He had a family history of both autism and ADHD. His older brother had autism and ID, his two cousins had language disorder (LD) and one uncle had severe ID and autism. At T1, Michael was diagnosed with Autistic Disorder and average intellectual functioning. He did not use more than a handful of words (PARIS level 2), scored low on RDLS receptive (10 raw score) and was very difficult to understand. He had difficulties with social interaction/communication such as turn taking and eye contact and seldom answered questions. Both parents were also well aware of his language problems, and had sought contact with the clinic on their own.

At the T2 assessment when he was 4.8 years old his parents reported that his major problems regarding language functioning and speech output still remained. They were concerned about his non-fluent staccato like speech, and he used plenty of iterations.

It is clear from Table 2, that this boy, in addition to his autism, also had distinct problems with language best described as ALD. He scored low on receptive language measures (TROG-2 and PPVT III), but spoke in 3-word utterances. He was still difficult to understand both for parents and others, and the LD had an impact on his everyday functioning. He only had 14 established consonants out of 18 in his repertoire and 4 phonological simplification processes.

At T3, at the age of 7.8 years Michael had been attending a special school unit for children with ALD since T2. He was still rather difficult to understand due to speech output problems, but managed to repeat 23 out of 30 one to five non-word syllables. The SLP also noted that he often, in his spontaneous speech, included one to three extra syllables in word final position, e.g., ([havet-te-te] in English “the sea”). The parents were concerned about this “stuttering” but the SLP interpreted it as problems with semantics/lexical skill and mazes (Bishop & Adams, 1989).

Indeed, his receptive language ability ranged from clearly below average on vocabulary (standard score 77 on PPVT-III) to the lower part of average on TROG 2 (standard score 88). Moreover, expressive grammar measured by recalling sentences (scaled score 5) was clearly below average. His low language abilities were also obvious on the BST, especially on the Information (18 raw scores) with a result corresponding to a child of 4.5 years, 6 Sentence Length (that corresponds to the result of a child of 4.9–5.8 years) and Subordinate Clauses 3 (corresponds to a result of child of 5.9–6.8 years of age) (Svensson & Tuominen-Eriksson, 2000).

His non-verbal skills measured with the matrices (T-score 63) showed a result clearly above average. Thus, at T3 Michael still had a clear profile of ASD and LD even though his parents rated him low on ASSQ, 10, which does not indicate ASD (since >18 is a suggested cut-off for ASD diagnosis).

“William” Case 3, (AGD)

“William” was 2.7 years old at his first assessment and showed great difficulties with compliance in the test situation. He had no family history of autism or DLD but fulfilled criteria for Autistic disorder and ID. At the T1 SLP assessment, he did not understand single words or simple instructions and scored 0 on RDLS receptive. He used a few single words and was rated as PARIS level 2, but most of his speech output was direct or delayed echolalia. He had severe difficulties with social communication and interaction and used very few gestures and mimic spontaneously. At age 4.9 years he came to the clinic for the T2 follow-up assessment. His diagnoses remained, and he exhibited severe problems within every area tested, i.e., autistic severity, language and cognitive functioning, but his behaviour had improved. He used a little bit more facial expressions, such as smiles and laughter when he was happy and he used more gestures spontaneously.

His parents used signs to support his speech as did some of the preschool staff. His language skills were still poor, he spoke in 1–2 word utterances and had severe problems with receptive language. He only had 10 established consonants out of 18 in his repertoire and 4 phonological simplification processes. At the T3 assessment, he was 7.3 years old and had started at a special school unit with one teacher per pupil. He mainly used single words or short phrases spontaneously and still had frequent echolalia in his speech output. His use of signs/gestures had also improved a little and the parents and school used picture support. William´s compliance in the testing were also improved to some extent but he could not participate in some of the tests chosen for the T3 follow-up. He scored significantly below average on receptive grammar (standard score 55 on TROG −2), receptive vocabulary (standard score 40 on PPVT_III) and on recalling sentences (scaled score 1 on CELF-4). His non-verbal functioning was also significantly below average (T-score 28 matrices WASI), indicating that William still had an AGD profile. His autism presentation, rated on ASSQ, was 20, which clearly indicates autism symptoms (since >18 is a suggested cut-off for ASD diagnosis).

“Linda” Case 4, (ALN)

“Linda” had an older brother who was diagnosed with ASD and ID at the CNC. During his assessment, the parents also raised concerns about Linda, even though she had completely different symptoms. She had no difficulties with language development, but had frequent tantrums and behavioural difficulties. At the T1 assessment at age 29 months, she used multiword utterances. Also, at the SLP assessment, she had one minor absence attack, which later was diagnosed as epilepsy. At 4.6 years of age, Linda came for her T2 assessment. Her language development had followed the expected trajectory and she had no problems with receptive or expressive speech and language (Table 2). Her autism severity score and subscores, were almost at the same level as at T1, and she was now diagnosed with Asperger´s syndrome. She had an age appropriate speech with 17 established consonants out of 18 in her repertoire and 0 phonological simplification processes. Linda could also name all letters in the Swedish alphabet at this age.

At T3, she was 6.9 years old and attended a regular school in a class with 26 pupils, and had no extra support in the classroom. She said “I love to read but Mathematics’ is even better” when the SLP asked her about the school. Her results showed a positive development on all tests (see Figure 1) even though the BST Information and Sentence Length (see Figure 2) indicated some problems with conveying story information, i.e. interpreting implicit information, and the Subordinate Clauses score indicated less complex expressive grammar. Her autism presentation, rated on ASSQ, was 31, which clearly indicates autism symptoms (since >18 is a suggested as a cut-off for ASD diagnosis).

“Tina” Case 5, (ALD)

“Tina” had a typical development during her first year of life. She came from a bilingual background Filipina (mother) and Swedish (father), with Swedish judged as her best language) and had no family history of autism or DLD. The parents described a language regression: Tina had used canonical babbling and single words as expected but then her language development stood still and she lost her early words. When Tina came to the CNC for her T1 assessment at 3 years of age, her parents told that she used 12 spoken words and some gestures (Table 1). She did not use any words at all during the assessment, had severe problems with language comprehension, but understood some spoken single words. The clinical multi-professional assessment revealed (PIQ = 86) and (ADOS sev = 6), resulting in two diagnoses – ASD and LD. At the age of 5.4 years (T2) she had attended preschool and had intervention/support for two years from the habilitation services. The parents reported some progress in her language development; she now used a few communicative sentences. She had an age-appropriate speech with 17 established consonants out of 18 in her repertoire and 2 phonological simplification processes (Table 3). Her comprehension of words and sentences were still clearly below average and impaired her daily life functioning both at preschool and at home.

Tina was 7.4 years at T3 and still had an ALD profile: she scored significantly below average on receptive language (standard score on TROG-2) (see Figure 1), receptive vocabulary (standard score on PPVT-III) and on Recalling Sentences (scaled score on CELF 4). She now spoke in 5-word long sentences (Figure 2), without any subordinate clauses, when retelling the Bus Story and had severe difficulties in remembering the story content. She also had severe difficulties on NWR which is considered a clinical marker for LD (Estes et al., 2007) and could only repeat 17 out of 30 non-words. Her non-verbal skills were still average (T-score 53). Her autism presentation, rated on ASSQ, was 20, clearly indicating autism symptoms (>18 is a suggested cut-off for ASD diagnosis).

“Anna” Case 6, (AGD)

“Anna” was 3.8 years at her first assessment. She had no family history of autism or DLD and the parents reported typical babbling and word development. Anna had older siblings and the parents became aware of her difficulties very early on. She was satisfied to be by herself, and not interested in sharing or showing her interests. She often played with small toy animals lining them up in long lines and sorting them depending on their colour. She fulfilled criteria for Autistic disorder and ID after the multidisciplinary assessment. (PIQ = 77) and (ADOS severity = 6). At the SLP´s T1 assessment, she talked a great deal, but mostly used echolalia and could also “read a story” from a well-known book verbatim, but did not understand any word or instructions in the receptive language test (Table 1). At T2, her receptive language was clearly below average but her vocabulary was average. She spoke in 3–4 word long sentences, had established almost all of the Swedish consonants and used two phonological simplification processes. Her autistic presentation was somewhat increased but her total adaptive functioning was identical to T1. At the third follow-up, conducted at age 7.3 years her language and word comprehension had decreased remarkably (figure 13) but she could repeat sentences below average and repeat non-words (18/30) on an average level. She had severe difficulties with story retelling (BST), especially in conveying story information and use of subordinate clauses (see Figure 2) with a result corresponding to a typical child of 3.9–4.8 years (Svensson & Tuominen-Eriksson, 2000). Her autism presentation, rated on ASSQ, was 31, clearly indicating autism symptoms (>18 is a suggested cut-off for ASD diagnosis).

In order to capture the children’s everyday communicative functioning according to the parents, we used the CCC-2. The results are presented and illustrated for girls and boys separately in Figures 3 and Figures 4.

Figure 3. Ten subdomains within the CCC-2 (scaled scores min-max 1–19, mean = 10) for the boys.

Figure 4. Ten subdomains within the CCC-2 (scaled scores min-max 1–19, mean = 10) for the girls.

Discussion

The present case study aimed to provide a broad and detailed description of language development and language skills in six children with ASD with or without accompanying LD and concurrent non-verbal cognitive disability. This is in accordance with the specification requirement of language and cognitive level in the DSM-5, which is one of the major systems used for diagnosing ASD (APA, 2013). The six children came from a general population-based sample and had screened positive for ASD during universal screening at age 2.5 years (Kantzer et al., 2013, 2018).

The first assessment of these children was performed around age 3 years, the second at age 5 years and the third when they were 8 years old. They were assigned a group based on their linguistic and non-verbal test performance results from the first T1 assessment and we were interested in how these subgroups developed over time by exploring language and cognitive profiles from each time point. We tried to take a strong developmental approach by describing their development in detail in order to present the well-known clinical linguistic heterogeneity among children with ASD (Boucher, 2012; Kjelgaard & Tager-Flusberg, 2001).

The key findings from this serial case study are that: (1) many children with ASD also have language difficulties irrespective of whether they also have non-verbal cognitive impairment or not, and (2) the subgrouping of children with ASD based on linguistic and non-verbal functioning was rather stable over time, and (3) upon closer examination of their test results their speech and language difficulties are rather similar to those seen in DLD.

In contrast to earlier studies, the parents included in our study did not report any atypical babbling for any of the children (Luyster et al., 2011; Oller et al., 1999; Paul et al., 2011). Only one girl with ALD had language regression with loss of words. According to Ozonoff et al. (2011) the vast majority of children who experience a regression lose behaviours related to social interest and engagement, such as eye contact and response to their names, but some children also lose spoken language, although this is less universal (Ozonoff et al., 2011). This language regression has also been demonstrated in children with typical development, but to a much lesser extent (Brignell et al., 2017).

As expected, our results showed that children with AGD performed lowest on a range of tests used in this study at different time points. Interestingly, in terms of language comprehension at T1, the two children with AGD performed lower than the children with ALD, but they used more spoken words. At T2, this discrepancy in language comprehension and production between AGD and ALD were evened out. Nevertheless, both groups have obvious language problems. When it comes to phonological ability measured as consonant inventory, one boy (AGD) had 10 and one boy (ALD) had 14 established consonants at age 5, the time at which the 18 consonants in the Swedish language should be mastered. The result of the receptive grammar was below 2 SD for both groups. At T3, the language problems in AGD and ALD remained. Thus, it seems that the only discrepancy between the two groups were in non-verbal cognitive performance. Furthermore, regarding autistic severity at 5 years of age, the ALN group presented with fewest symptoms while the other two groups had a more equal result.

The visual presentations of the CCC-2 collected at T3 revealed that the parents rated their children rather low on all measures. For example, Speech and Syntax were rated unexpectedly low irrespective of their child’s subgroup belongingness, while the communication measures (Coherence, Initiation, Use of Context, Social Communication and Social relations) were rated low as expected.

Based on our results showing different patterns based on linguistic and cognitive performance, it is of great importance that children with ASD are assessed and thoroughly investigated with regard to speech and language ability (Carlsson, 2019). It is also important to take parents’ opinions into account in order to be able to give the child and the family appropriate and well-timed intervention and support (Andersson et al., 2014). Since the results show a rather stable path through preschool and early school age, it seems important to have an early plan for SLP intervention and follow-up. Children with ASD in isolation, that is ASD “only” is quite rare and these children often function relatively well in their daily lives, while children that have ASD combined with other difficulties such as LD or ID, i.e. Autism +, are impaired to a greater extent and tend to need more support (C. Gillberg & Fernell, 2014). Consequently, it is important to acknowledge that many children with ASD have language difficulties that might hinder their development and they should be entitled to SLP support and intervention.

Limitations

A case study – even though it presents data from true cases at three time points – has limitations when it comes to generalisability. Even so, we took great effort in defining the subgroups of children by using theoretical-based cut-offs of both language and non-verbal cognitive level in order to show the great heterogeneity in children with ASD. The discrepancy between receptive and expressive language performance at different time points would of course be very interesting to investigate in a much larger sample, which is underway.

To summarise, this study presents a detailed and in-depth knowledge of six individuals with different presentations with regard to language and non-verbal skills. The findings showed that children diagnosed with ASD and LD at 3 years of age also have persistent language difficulties four to five years later. Based on this, further research should be carried out to explore how language skills develop over a longer time, by focusing on older children and adolescents. Another possible area of future research would be to investigate and evaluate specific intervention and support for children with ASD and LD; we need to explore the effectiveness and evidence for such interventions methods for children with ASD.

Declaration of interest

No potential competing interest was reported by the authors.

Correction Statement

These authors contributed equally to this work.This article has been corrected with minor changes. These changes do not impact the academic content of the article.

Additional information

Funding

This work was supported by the Swedish Research Council for Health, Working Life and Welfare [project No. 2013-00092].

Appendix

Appendix Table 1. Outcome measures and materials used within the study at time point 1, 2 and 3.

Outcome measures	Test materials	T1	T2	T3
Receptive language	Test of Reception of Grammar-2 (TROG-2) (D Bishop, 2003) (Bishop, 2009)		X	X
	Reynell Developmental Language Scales III (RDLS) (Edwards et al., 1997) Swedish norms (Eriksson. L. & Grundström, 2000; Lindström & Åström, 2000)	X
Vocabulary	Peabody Picture Vocabulary Test III (PPVT III) (Dunn & Dunn, 1997)	X	X	X
	Swedish Early Communicative Development Inventories (SECDI):  Words and gestures (Eriksson & Berglund, 1999; Fenson et al., 1993)	X
Expressive language	The PARIS scale (Philippe et al., 1999)	X	X	X
Sentence repetition	Recalling Sentences (CELF-4) (Semel, Wiig, & Secord, 2003)			X
Oral non-word repetition	Non-word repetition (Radeborg, Barthelom, Sjöberg, & Sahlén, 2006)			X
Narrative ability	The Bus Story Test (BST) (Renfrew, 1997), (Svensson & Tuominen-Eriksson, 2000).		X	X
Pragmatic/communicative functioning (parent-rated)	Children’s Communication Checklist- second edition (CCC-2) (D. Bishop, 2003)			X
Autism severity & symptomatology	Autism Spectrum Screening Questionnaire (ASSQ) (Ehlers, Gillberg, & Wing, 1999)			X
	The Autism Diagnostic Observation Scale Generic (ADOS) (Lord et al., 2000)	X	X
Adaptive behaviors	Vineland Adaptive Behaviors Scales (VABS), (Sparrow, Cicchetti, & Balla, 2005)	X	X
Non-verbal cognitive ability	WASI sub-test Matrices (Wechsler, 1999)			X
	Griffiths’ Developmental Scales (Alin-Åkerman & Norberg, 1991)	X	X
Non-verbal sequential reasoning	Picture Arrangement (WISC III) (Wechsler, 1991)			X