Prevalence and explanation of co-occurrence of other disorders with SLI
Supervisor: Dr. Margriet J.M. Heim Second reader: Dr. Jan de Jong
Master: General Linguistics – Clinical Linguistics
Acknowledgements
The completion of this thesis would not have been possible without the help of many people. First of all, I would like to thank my two supervisors dr. Jan de Jong and dr. Margriet Heim. Jan, thank you for the beginning. You helped me in shaping my ideas and gave me a good start with my thesis. Your enthusiasm was much appreciated. I would also like to thank you for being my second reader. Margriet, thank you for the last part of my journey. I know I was not a perfect student, but you stuck with me anyway. Thank you for your patience and your perseverance. Without you I would still be lost. Moreover, I would also like to thank dr. Judith Rispens, who inspired me to conduct research about the co-occurrence of other disorders with SLI.
In addition, I would also like to thank my family for being there for me and for believing in me no matter what. Pap, thank you for your patience with me and my statistical problems. Mam, you are my rock. I can always depend on your positivity to help me get through the rough patches I encountered during my writing. I love you both. Furthermore, I would like to thank my best friend, Nienke. Thank you for always believing in me and in my abilities. Now I can finally say “we did it” too. Thank you for reading everything I wrote and giving me feedback were I needed it. Thank you for taking this journey with me and bearing with me, even if I was a bit cranky at times.
I would also like to thank my colleagues for supporting me during this period, especially Elles. Elles you were my sounding board for new ideas, even when you did not know what I was talking about. You were always positive when I was down and you would always make me laugh, when I was crying. Thank you for being a friend. In addition, I would like to thank Orion College West for giving me the opportunity to go through her archives and letting me use her data. Thank you, Audrey Franssen and Marian Hoefnagel, for the help during my thesis and my internship.
A final thanks to all my friends and family who never stopped believing in me. Opa, Oma, Joris, Eva, Maarten, Tony, Nienke L., Soraya, Corinne, Jeannette, Eline, Emma, Lotte, Martine, the regulars at Kwekkeboom and everybody I am forgetting! Thank you. Thank you all for taking this journey with me. It was a long and hard experience for me, but I do believe that it has made me a stronger and more complete person at the end. It has thought me to never stop believing in myself and that if you want something bad enough and if you are willing to work hard for it, you will always succeed. I am proud of what I have accomplice and therefore I dedicate this thesis to myself.
Abstract
This thesis aims to answer the following questions: (1) Which disorders co-occur with SLI?
(2) Can these co-occurrences be explained by a common underlying deficit?
(3) What are the frequencies of these co-occurrences in a Dutch data set of children with SLI?
To answer these questions, a literature review and a small desk study were carried out. Both the literature review and the desk study showed that ADHD, dyslexia and ASD often co-occur with SLI (question 1). To explain these co-occurrences two models were proposed in the literature review: the Multiple Deficit Model and the Comorbidity Model. The Multiple Deficit Model claims that there could be an overlap in the underlying deficit that caused SLI and a co-occurring disorder. The Comorbidity model claims on the other hand that there is no overlapping underlying deficit which can explain the co-occurrence of SLI with other disorders. The literature review showed that the studies conducted on this topic are inconclusive and that more research is needed to entangle the precise relationships between the co-occurring disorders.
After the literature review, a desk study was conducted. For this study, school records of 74 participants of the Orion College West were analysed. These participants were selected based on the basis of their SLI cluster 2 admission. Subsequently, information on the co-occurrence of other disorders was gathered from these individual records and the frequencies of these co-occurrences were counted. However, the last group may even have been larger due to the fact that applying for an ASD cluster 2 admission is less time consuming. The children with an ASD cluster 2 admission were not selected for this study, but may still have SLI as well.
With regard to the first research question, which disorders co-occur with SLI, the same disorders as described in the literature were found, such as: ADHD, dyslexia and ASD. With regard to the second research question, whether the found occurrences can be explained by a common underlying deficit, no conclusive answer was found.
There were some answers found though to the third research question, what the frequencies are of the found co-occurrences in a Dutch data set of children with SLI. Results from the desk study show that the group of children with only SLI is the largest followed by the group of children with SLI and dyslexia and SLI and ASD.
The implications of the findings of this study are that if a child, for instance, has SLI, there is a substantive chance that this child also has dyslexia or ASD or even both. Of course, this is not always
the case, but teachers, remedial teachers and speech therapists should be more aware of the possibility. This in turn might help with an earlier diagnosis, so more suitable help can be offered to the child at a younger age.
Moreover, as no conclusive answer was found to the second research question, more research is necessary to investigate whether co-occurrences of other disorders with SLI can be explained by a common underlying deficit. A possible reason that this research question could not be answered is that the data which was gathered for this research was often incomplete and the children involved should have been (re)tested extensively. For several reasons this was not feasible. Therefore a research proposal was written to direct future research. This proposal gives guidelines for accurate testing of children in order to find answer to the question of a possible underlying deficit.
Table of Content
Acknowledgements...2
Abstract...3
Table of Content...5
1 Introduction...7
2 Different Disorders that Co-occur in Children with SLI According to the Literature...9
2.1 Specific Language Disorder (SLI)...9
2.2 Attention Deficit Hyperactivity Disorder (ADHD)...9
2.2.1 Symptoms and Epidemiology...9
2.2.2 The Underlying Deficit of ADHD...10
2.3 Dyslexia...11
2.3.1 Symptoms and Epidemiology...11
2.3.2 The Underlying Deficit of Dyslexia...11
2.4 Autism Spectrum Disorder (ASD)...13
2.4.1 Symptoms and Epidemiology...13
2.4.2 The Underlying Deficit of ASD...13
3 Explanations for the Co-occurrence of SLI with Other Developmental Disorders...15
3.1 The Multiple Deficits Model and the Comorbidity Model...15
3.2 Common Cause?...16
3.2.1 SLI and ADHD...16
3.2.3 SLI and ASD...21
3.2.4 Conclusion...23
4 The Current Study...25
4.1 Introduction...25
4.2 Participants...25
4.3 Data Collection and Analysis...26
4.4 Reliability and Validity...26
5 Results and Conclusion...28
5.1 Co-occurrences...28
5.2 Conclusion...28
6 Discussion and Future Directions...31
6.1 Research Proposal...31 6.1.1 Participants...31 6.1.2 Design...32 6.1.3 Procedure...33 6.1.4 Data Analysis...33 References...35 Appendix I...39
1
Introduction
In general, children learn their language without too much trouble, but some children have problems acquiring their first language. These problems can have multiple reasons, for example: language deprivation, attention problems (Geurts & Embrechts, 2008), other disorders or they may be the result of a primary (specific) language disorder. These language disorders manifest themselves in the language production of an individual.
A distinction can be made in primary and secondary language disorders. Secondary language disorders are the result of another disorder, like Attention Deficit Hyperactivity Disorder (ADHD) or Autism Spectrum Disorder (ASD) (Williams & Lind, 2013). This means that for instance the disorder ADHD besides causing attention problems also causes language problems in one individual. On the other hand, language disorders that are not secondary to another disorder, are called primary language disorders, in which language lags behind other aspects of development.
A term often used for primary language disorders is Specific Language Impairment (SLI). This neurodevelopmental disorder affects between 3 to 10% of all children (Bishop & Snowling, 2004). The language of a child with SLI is significantly below the level expected based on age and IQ without an apparent reason (Bishop, 2006). This means that there is a problem in language comprehension, language production or in both. For example, a seven-year-old child with SLI might sound like a four-year-old typically developing child.
SLI is not a clearly delimited single disorder. The disorder includes people with varying profiles arising from (combinations of) deficits in phonology, morphology, syntax, semantics and pragmatics (Vandewalle, Boets, Ghesquière & Zink, 2010). Different combinations of these deficits can cause either receptive or expressive difficulties. This can lead to the manifestation of different features, like a delayed onset of speech, immature or deviant production of speech sounds, use of simple grammatical structures, limited vocabulary and difficulties in understanding complex language, especially when spoken rapidly (Bishop, 2006, p. 218). Further information on SLI will be given in section 2.1.
Individuals with SLI often have other developmental disorders as well. In a preliminary literature search dyslexia, ADHD and ASD are found as often co-occurring with SLI (information on these disorders can be found in chapter 2). Often the distinctions between these different disorders are not clear cut (Bishop & Snowling, 2004, Bishop, 2010, Hill, 2001). Moreover, many disorders may occur at the same time (Hill, 2001) and have a shared etiology (Bishop, 2010). However, the reason as to why these disorders co-occur so often together is still an area of debate. Some researchers say that there is an overlap in underlying deficits that cause the different disorders, while others say that there is no
overlap at all. For example, Bishop & Snowling (2004) claim that one of the underlying deficits causing both dyslexia and SLI is a deficit in phonological processing. Catts, Adlof, Hogan & Weismer (2005), on the other hand, say that there is no overlap in the underlying deficit. As these findings are contradicting each other, this thesis therefore tries to fill this gap in the literature by aiming to answer the following questions:
(1) Which disorders co-occur with SLI?
(2) Can these co-occurrences be explained by a common underlying deficit?
(3) What are the frequencies of these co-occurrences in a Dutch data set of children with SLI?
In the following chapter two the different disorders that co-occur with SLI will be discussed. These disorders were found with a preliminary literature search. The disorders that will be discussed are ADHD, dyslexia and ASD. The symptoms and the epidemiology of the disorders will be discussed, as well as underlying deficits that may cause the disorders. In chapter three possible models that may explain the relationship between SLI and the other disorders will be addressed. Moreover, in this chapter it will be discussed whether these co-occurrences can be explained by a common underlying deficit (research question two) based on results of the literature review. After this theoretical framework chapter four will investigate such co-occurrences within a specific Dutch data set on the basis of a desk study. The identification of disorders that co-occur in these data and the frequencies of these co-occurrences are distracted from data which have been collected from a secondary school for special education for children with language problems in the Netherlands; the Orion College West. Chapter five will discuss the results from this desk study and provide a conclusion. And finally, chapter six will discuss the found results and give recommendations and directions for future research.
2
Different Disorders that Co-occur in Children with SLI According to the Literature
This chapter will elaborate on SLI and different disorders that co-occur with it. The first section will give a description of the symptoms and epidemiology of SLI (2.1). Subsequently, three disorders which were found to co-occur frequently with SLI will be discussed: Attention Deficit Hyperactivity Disorder (ADHD) in paragraph 2.2, dyslexia in paragraph 2.3 and Autism Spectrum Disorders (ASD) in paragraph 2.4. Moreover, in each of these paragraphs, a possible underlying deficit will be discussed. 2.1 Specific Language Disorder (SLI)
The problems that are caused by SLI often manifest themselves in the area of morphosyntax (Catts et al., 2005). Although SLI mainly manifests itself in the spoken language, it seems that it can also lead to problems in reading and writing (DeBree, Rispens & Gerrits, 2007).
SLI is not a clearly delimited single disorder. There are varying profiles with deficits in different areas of the language (phonology, morphology, syntax, etc.). These deficits in the different parts of communication can either be receptive, expressive or both.
The diagnosis of SLI is based on excluding other possible causes for language impairments, such as hearing- or vision problems or whether the non-verbal intelligence of an individual falls within broadly normal limits (Bishop, 2006). Hence, in order to be diagnosed with SLI, an individual needs to have no known neurological, physical, emotional, social problems or language deprivation (McArthur, Hogben, Edwards, Heath & Mengler, 2000).
Underlying deficits that may cause SLI are thought to be found somewhere in the working memory (Whitehouse, Barry & Bishop, 2008). For instance, children with SLI often have problems with their verbal short-term memory (vSTM). The vSTM is often tested with tasks that test the ability to retain auditory information, like non-word repetition tasks (NWR), which will be explained in section 2.3.2. Children with SLI are often able to repeat short non-words, but seem to fail in accuracy and speed when the length of the non-words increase. This suggests that the problem lies in holding phonological information in the memory, rather than in perception and production.
2.2 Attention Deficit Hyperactivity Disorder (ADHD)
2.2.1 Symptoms and Epidemiology
Attention Deficit Hyperactivity Disorder (ADHD) is a neurodevelopmental disorder that is estimated to have a prevalence of 3 to 11% depending on the population studied and the measures used (Williams, Stott, Goodyer & Sahakian, p. 368, 2000). The main features of the disorder are
inattention, impulsivity and hyperactivity. Some other characteristics of ADHD are poor self-regulation, planning, execution and monitoring of own behaviour (Hutchinson, Bavin, Efron & Sciberras, 2012). The disorder might also involve language problems, although language related impairments are not necessary to fulfil the criteria for ADHD (Geurts & Embrechts, 2008).
ADHD is, like SLI, a heterogeneous disorder. The DSM-IV therefore specifies three subtypes of ADHD based on the level of presenting symptoms (Jonsdottir, Bouma, Sergeant & Scherder, 2005). The first one is the inattentive subtype (ADHD-I), also known as attention deficit disorder (ADD), the second subtype is the hyperactive-impulsive type (ADHD-HI) and the last one is the combined subtype (ADHD-C) in which symptoms of the two subtypes are present.
Quite often children with ADHD have pragmatic language problems (Geurts & Embrechts, 2008). They show more stereotyped conversations, more problems with conversational rapport and more problems with social relationships compared to typically developing children (Bishop & Baird, 2001; In: Geurts & Embracht, 2008).
2.2.2 The Underlying Deficit of ADHD
One of the deficits that seems to underlie ADHD is a deficit in working memory (Martinussen & Tannock, 2006). Working memory (WM) is the ability to store and process information simultaneously for a brief period (Alloway, Rajendran & Archibald, 2009). WM can be specified using the model by Baddely and Hitch (1974) and Baddely (2010). This model is depicted in figure 2.1.
As can be seen in figure 2.1, the Central Executive (CE) supervises the subordinate components: the Visuo-spatial Sketchpad (VP), the Episodic Buffer (EB) and the Phonological Loop (PL). The CE holds and manipulates information that comes from the other components. The VP processes visual information and is also known as the visuo-spatial WM. The EB integrates visual and verbal information and creates a link to the long-term memory and the PL processes incoming auditory information, this element is also known as the verbal WM (Hutchinson et al., 2012). Together these elements form the working memory.
Different studies have been done as to where the possible underlying deficit is located in the WM (Hutchinson et al. 2012, Jonsdottir et al. 2005 and Alloway et al., 2009). However, these different studies often disagree on the location of the deficit in the WM. Some of these studies even disagree on a deficit in the WM being the possible cause of ADHD. This will be discussed further in 3.2.1. 2.3 Dyslexia
2.3.1 Symptoms and Epidemiology
Dyslexia is diagnosed when a child has severe problems in reading and spelling for no apparent reason. These difficulties in reading and spelling are persistent to the usual didactic measures and remedial efforts (Vandewalle, Boets, Ghesquière & Zink, 2012). People with dyslexia often have poor reading skills, illegible handwriting and a tendency to read, write and/or speak letters and words in another sequence (Dyslexia, In: Encyclopædia Britannica, 2013). Dyslexia is an heterogeneous disorder. This means that the behavioural manifestations and the severity of the disorder vary per person.
Like SLI, dyslexia is diagnosed based on exclusion criteria. People with dyslexia should have normal hearing and adequate intelligence. Moreover, there should be no other neurological, physical, emotional, social problems or language deprivation that can explain the disorder.
Dyslexia affects between 3 to 10% of the population (Bishop & Snowling, 2004). Moreover, genetics seem to play an important role in the chance of a child being affected by dyslexia. When a first degree relative has dyslexia, the risk of the child developing the disorder is estimated around 40% (De Bree, Rispens & Gerrits, 2007).
2.3.2 The Underlying Deficit of Dyslexia
There is still some debate on the cause of dyslexia. However, most of the present data point towards an underlying deficit in phonological processing (Vandewalle et al., 2012). This underlying deficit in phonological processing is explained in the phonological deficit theory. According to this theory, phonological processing consists of three distinct but related components: phonological awareness,
verbal short-term memory and lexical access. These three components are used in the representation, manipulation, short-term storage and retrieval of speech sounds.
First of all, Phonological Awareness (PA) is the conscious sensitivity that typically developing children develop to the sound structure of language (Vandewalle et al., 2012). This ability allows us to analyse words into consonants and vowel segments (Rispens, 2004). PA is used when learning how to read, since the sounds need to be mapped onto the right corresponding graphemes. PA is measured with tasks that test the ability to distinguish and manipulate syllables or phonemes, such as phoneme deletion or spoonerism tasks (Vandewalle et al., 2012).
Second, Verbal Short-Term Memory (vSTM) is the ability to maintain auditory information for a short period of time (Vandewalle et al., 2012). To maintain this information the phonological loop is used. The auditory information is stored as a phonological trace for approximately two seconds, after which it is reactivated by articulatory rehearsal and is allowed to be stored again for a short period of time as a phonological trace (Rispens, 2004). In other words, the phonological loop makes it possible to keep the auditory information activated by reactivation; hence the term loop. The vSTM is tested with tasks that test the ability to retain auditory information, like the digit span task and the Non-Word Repetition task (NWR) (Vandewalle et al., 2012).
And third, lexical access is the ability to retrieve lexical sound-based representations from the long-term memory (Vandewalle et al., 2012). These representations that need to be retrieved are formed via auditory perception. From the different tokens of one representation of one entity that are heard, one phonological representation is formed and then placed in the long-term memory (Rispens, 2004). For example, different forms of [p] (tokens) are heard. Together these tokens form one phonological representation of /p/ which is placed in the long-term memory. Lexical access is tested with Rapid Automatized Naming (RAN), like letter, picture or colour naming (Vandewalle et al., 2012).
All three components, PA, vSTM and lexical access, are of importance when learning how to read. Therefore PA and lexical access are considered to be good predictors of how a child will perform when learning to read. Moreover, PA can be considered to be of greater importance in the first stages of learning, whereas the performance on RAN is more important in the later stages, when a child learns to read more fluently and learns to comprehend what he is reading (Vandewalle et al., 2012). An individual with dyslexia is likely to score low on tests of all components. However, due the heterogeneity of the disorder, the severity of the deficits in the different components may differ per person. Different studies have been conducted to investigate whether a deficit in these components can also be one of the underlying deficits which causes SLI (see also paragraph 3.2.2). Studies which investigated this include Vandewalle et al. (2012), Bishop, McDonald, Bird & Hayiou-Thomas (2009) and Catts et al. (2005).
2.4 Autism Spectrum Disorder (ASD)
2.4.1 Symptoms and Epidemiology
Autism Spectrum Disorder (ASD) is an umbrella term for four specific diagnoses: (1) autistic disorder, (2) Asperger’s disorder, (3) childhood disintegrative disorder and (4) pervasive developmental disorder not otherwise specified (PDD-NOS) (DSM-V, 2013). Early studies estimate the rate of children with ASD to be two to five children per 10 000, but more recent studies report a higher rate, namely 60 children per 10 000 (Frith & Happé, 2005). Volkmar & Pauls (2003) explain this increase in frequency by differences in methods used to diagnose ASD, growing public awareness and changes in the definition of ASD. They claim that at least one in 1000 children shows signs of autism. Furthermore, even a larger number of children do show symptoms of ASD, but without meeting the full criteria for a diagnosis (p. 1134).
ASD is characterised by severe difficulties in social interaction, communication and unusual behaviours (Volkmar & Pauls, 2003). For example, people with ASD may respond inappropriately in conversations or interpret utterances literally and fail to notice sarcasm or figurative speech (Frith & Happé, 2005). Other symptoms include difficulties with change in either environment or routines, problems with making friends, an intense focus on a certain subject and stereotyped movements (Volkmar & Pauls, 2003).
ASD is, like SLI, ADHD and dyslexia, a heterogeneous disorder. Which means that people with ASD might have different symptoms and that the severity of these symptoms differ from mild to severe (DSM-V, 2013).
2.4.2 The Underlying Deficit of ASD
Since the group of people with ASD is so diverse, it is difficult to find a possible underlying deficit that might also underlie SLI. This is why, for the purpose of this thesis, the group will be narrowed down to individuals with ASD that have language difficulties. Some people with ASD have normal language skills, while others with ASD have language skills comparable to the language skills of children with SLI. The latter group is often referred to as ALI (Autism Spectrum Disorder with language impairment) (Loucast, Riches, Baird, Pickles, Simonoff, Chandler & Charman, 2013).
The observed language impairment in ALI is comparable to that of SLI. Therefore it is often thought that there might be an overlapping in language phenotypes (Riches, Loucast, Baird, Charman & Simonoff, 2011). Moreover, because of these similarities it is thought that the language problems observed in ALI might also be caused by an underlying deficit in vSTM. If the language problems of individuals with ALI are similar to those of individuals with SLI, the problems found, with for instance non-word repetition (mentioned in section 2.1), will also be found in children with ALI. Many studies
have been conducted to find out whether this overlap between ALI and SLI exists (see also paragraph 3.2.3), including Taylor, Maybery, Graynder & Whitehouse (2014), Riches et al. (2011) and Whitehouse et al. (2008).
3
Explanations for the Co-occurrence of SLI with Other Developmental Disorders
In the past, different models have been developed to help explain the co-occurrences of SLI with other disorders. In section 3.1, two of these models will be discussed, namely the Multiple Deficits Model (adaption of the model by Catts et al., 2005) and the Comorbidity Model (adaption of the model by Catts et al., 2005). As will be explained, these models contradict each other in their explanation of the co-occurrence of SLI with other disorders. That there is no consensus regarding this topic is also further demonstrated by research. Some researchers believe that there might be an overlap in underlying deficits that cause the different disorders, while others think that there is no overlap at all. The following section, 3.2, will discuss different studies that have been conducted on this possible overlap of underlying deficits in SLI and the different disorders; ADHD, dyslexia and ASD. 3.1 The Multiple Deficits Model and the Comorbidity Model
The Multiple Deficits Model and the Comorbidity Model share both the assumption that different underlying deficits can cause one type of disorder and that multiple disorders can co-occur in one individual. However, the two models differ on one key component, namely on how these co-occurrences of different disorders in an individual are explained.
The first model, the Multiple Deficits Model, claims that there can be an overlap between the two co-occurring disorders (see figure 3.1).
Figure 3.1 Multiple Deficits Model (Adaption of model by Catts et al., 2005).
Figure 3.1 shows that two disorders, A and B, share an underlying deficit X and that it is possible that these disorders result in a similar symptom R. However, the disorders are distinct entities even though they share an underlying deficit, because they do not share all underlying deficits (W and Y); there is an overlap (Bishop & Snowling, 2004). For example, the underlying deficit X that causes
dyslexia (disorder A) could be a phonological processing deficit. This deficit also causes SLI (disorder B) together with other cognitive deficits (underlying deficits Y) (Bishop & Snowling, 2004).
The second model, the Comorbidity Model, claims that the different disorders do not share any underlying deficits and that the disorders are therefore entirely distinct entities. The two disorders can, however, co-occur in one individual, making the two disorders true forms of comorbidity. This can be seen in figure 3.2.
Figure 3.2 Comorbidity Model (Adaption of model by Catts et al., 2005)
Figure 3.2 shows that the disorders do not only share any underlying deficits, they also do not share any behavioural symptoms.
The Comorbidity Model and the Multiple Deficits Model agree on the fact that different disorders can co-occur in one individual. However, the Multiple Deficits Model claims that there is an overlap in the deficits that causes the disorders, whereas the Comorbidity Model claims that there is no common ground between the co-occurring disorders. The rest of this chapter will discuss which of the models may explain the different co-occurrences of disorders with SLI best. This is done by discussing existing research on the credibility of an overlap in underlying deficits.
3.2 Common Cause?
3.2.1 SLI and ADHD
ADHD and SLI initially do not seem to have much in common. However, there seems to be an overlap in certain areas. One of these overlaps seems to be a working memory deficit in the cognitive profiles of ADHD and SLI. To specify the WM profiles of ADHD and SLI some studies have been conducted, such as studies by Jonsdottir et al. (2005), by Alloway et al. (2009) and Martinussen & Tannock (2006).
Jonsdottir et al. (2005) examined the impact of comorbid SLI on verbal and visuo-spatial WM in children diagnosed with ADHD-C. In their study children with ADHD-C with and without SLI and typically developing children were compared by measuring verbal and visuo-spatial WM. In total 49 children were tested within the age range of 8;6 and 12;6 years; 19 children were in the ADHD-SLI group, 15 children in the ADHD-non SLI group and 15 children in the typically developing group. The typically developing children were matched with the other participants on nonverbal intelligence. To measure the abilities of the children in verbal and visuo-spatial WM, different subtests from the
K-ABC1 were used. The subtests Number Recall and Word Order were used to determine the children’s
abilities on verbal WM. Moreover, Spatial Memory and Hand Movement were used to determine their abilities on visuo-spatial WM.
This study found no overlap in the possible underlying deficit that causes both SLI and ADHD. Jonsdottir et al. (2005) only found a deficit in the WM in the ADHD+SLI group, but they did not find the same deficit in the ADHD only group. The children with ADHD and SLI scored significantly lower on both Number Recall and Word Order (Jonsdottir et al., 2005). This would suggest that the deficit in WM found in children with both AHDH and SLI is caused by SLI and not by ADHD and that there is no overlap in the underlying deficit that might cause both disorders. This study, therefore, is more in line with the Comorbidity Model than it is in line with the Multiple Deficits Model.
Furtermore, Alloway et al. (2009) compared the WM profiles of children with different developmental disorders, namely SLI, Developmental Coordination Disorder, ADHD and Asperger syndrome. For the purpose of this thesis only the results on SLI and ADHD will be discussed. In their study 15 children with SLI and 83 children with ADHD-C were tested. Alloway et al. used the computerized and standardized tool Automated Working Memory Assessment2 to measure verbal
and visuo-spatial aspects of short-term memory and WM. Verbal and visuo-spatial WM were measured with tasks that involved both storage and processing of information, where verbal and viso-spatial short-term memory were measured with tasks that only involved the storage of information. The verbal aspect of the short-term memory was tested with recall sequences of verbal material. Moreover, the visuo-spatial short-term memory was tested with the presentation and recall of materials in a visual matrix. To assess the CE of the working memory more complex tasks were made, in which the amount of information to process and to store increased until errors were made. Alloway et al. (2009) found an overlap in WM profiles of children with SLI and children with ADHD and that these children both have a deficit in verbal WM. They also found that children with ADHD do not only have a deficit in the verbal WM, but also in the visuo-spatial WM. This is more in line with the idea that a deficit in the WM is a fundamental impairment in ADHD. Since they found an overlap
in the underlying deficit that might cause both disorders, this study is as opposed to the study of Jonsdottir et al. (2005) more in line with the Multiple Deficits Model than it is in line with the Comorbidity Model.
Finally, Martinussen & Tannock (2006) compared three clinical groups and a comparison group on their WM profiles. The four groups were children with only ADHD (62 children, all subtypes of ADHD), with reading disorder (RD) and language impairment (LI) (15 children), with both ADHD and RD/LI3 (32 children) and a comparison group with typically developing children (34 children). All children were aged between 7 and 13 years and spoke English as a first language. The comparison group was matched on age. Martinussen & Tannock tested in the domain of verbal WM, visuo-spatial WM and the CE, which they divided into a verbal and a spatial component. The verbal WM and CE were tested with the Digit Froward and the Digit Backward subtests from the WISC-III4. The spatial WM and CE
were tested with two brief tasks in which it was required to remember and reproduce a sequence of spatial locations in either the same or the reversed order of representation. The VS was also tested with the Finger Windows task form the Wide Range Assessment of Memory and Learning5. To tap into
the spatial CE the same task was used only in reverse.
Martinussen & Tannock (2006) found an overlap in the visuo-spatial WM and the CE6. They also found a deficit in the verbal WM of the group of children with RD/LI. This might indicate that deficits in this area of the WM are specifically associated with weaknesses in language and reading. Due to the overlap found in possible underlying deficits, this study is just like the study by Alloway et al. (2009) more in line with the Multiple Deficits Model.
The three studies discussed above all show different results. Jonsdottir et al. (2005) found no overlap in a possible underlying deficit causing both disorders (or even a possible underlying cause for ADHD). Whereas, Alloway et al. (2009) and Martinussen & Tannock (2006) did find an overlap. However, Alloway et al. and Martinussen & Tannock found the possible overlap in different areas of the WM: Alloway et al. (2009) found the deficit in the verbal WM while Martinussen & Tannock (2006) found it in the visio-spatial WM and the CE.
Based on the studies presented here and other studies that have been conducted so far on the subject of the co-occurrence of ADHD and SLI7, no conclusive answer can be given as to which model is best supported. The results differ too much; it cannot even be said conclusively if WM deficit is
3 These children were not diagnosed with SLI. The children were tested with the Clinical Evaluation of
Language Fundamentals – Third edition. To participate the children had to score 1.5 SD below the mean for age in either Expressive or Receptive Language scores or 1 SD below the mean on both Expressive and Receptive scores (Martinussen & Tannock, 2006).
4 Wechsler Intelligence Scale for Children – Third Edition, Wechsler, 1999. 5 Wide Range Achievement Test-3, Wilkinson, 1993.
6 When comparing the results of Alloway et al. (2009) and Martinussen & Tannock (2006) we have to keep in mind that Martinussen & Tannock (2006) use a RD/LI group instead of a SLI group.
really part of the underlying deficit that causes ADHD. It might be that the WM improves as soon as the subject uses medicine to control the ADHD, in which case the poor WM skills are a reflection of a lack of behavioural inhibition instead of a WM deficit (Alloway et al., 2009). The problem mainly lies in the fact that the tasks that are used to test certain areas of the WM often also tap into other areas of the WM. For example, the Hand Movement task Jonsdottir et al. (2005) used in their study correlated with the verbal WM as well as the visuo-spatial WM, but it was used to measure the ability of the latter. Another reason for these differences in findings are ascribed to dissimilarities in the groups that are examined and differences in diagnostic criteria and measures used to profile the WM. For example, the way that the control group is matched to the ADHD group might have an effect on the outcome, as well as the use of different subtypes of ADHD (Jonsdottir et al., 2005). To be able to give a conclusive answer more research is needed.
3.2.2 SLI and Dyslexia
When looking at the possible cause of dyslexia, a phonological processing deficit (Vandewalle et al., 2012), one might argue that this might also be one of the causes of SLI. However, not everyone agrees on this. Different studies have been conducted on this topic, two of which will be discussed in detail below, namely the study by Catts at al. (2005) and the study by Bishop, McDonald, Bird & Hayiou-Thomas (2009).
Catts et al. (2005) tested typically developing children, children with only SLI, children with only dyslexia and children with both dyslexia and SLI to investigate whether there is a difference in the phonological processing deficit. The children were tested over a long period of time, namely from kindergarten through to the eighth grade. Catts et al. were sceptical about the idea that this one deficit could cause both dyslexia and SLI. The PA and the vSTM of the children were tested with syllable/phoneme deletion tasks and non-word repetition tasks.
The study conducted by Catts et al. (2005) showed that the children with only dyslexia and the children with both dyslexia and SLI scored poorly on PA and NWR across the grades, whereas children with only SLI scored better on these tasks. However, when the children with SLI were compared to the typically developing children, the data showed that the children with SLI did score lower on the different tasks, although not statistically significant in all cases. These differences were mainly found in the early grades. This might indicate that children with only SLI have a mild deficit in phonological processing and that the deficit in phonological processing is more closely associated with dyslexia than it is with SLI (Catts et al., 2005). Therefore, according to Catts et al. a deficit in phonological processing cannot be the underlying deficit of SLI. Consequently, there is no overlap in underlying deficits. The research conducted by Catts et al., therefore, is more in line with the Comorbidity Model than it is with the Multiple Deficits Model.
Bishop et al. (2009) published their work as a reaction on the study published by Catts et al. (2005). They wanted to find out how it was possible that Catts et al. (2005) found that there were children with SLI who were competent in single word reading (RAN). Bishop et al. found these results particular and therefore wondered if the comprehension of these children with SLI is disproportionately poor. Next to that, they wanted to know if it was true that children with SLI did not have a deficit in phonological processing. In their study, Bishop et al. used same-sex twins aged between 9 and 10 years who have been selected by oversampling children at risk of language or literacy problems. Previous data were present on some of the children (aged 4 and 6). The children were divided into four groups: typically developing, SLI, dyslexia and SLI and dyslexia. The children were assessed with a number of tests to measure their language, literacy and nonverbal ability, like NWR, tapping sequence and MCSA Verbal Memory for Words and Sentences (Bishop et al., 2009). The results of the study conducted by Bishop et al. (2009) showed that SLI had a greater impact on reading comprehension than on single-word recognition. This is supported by the results that the children with SLI did score better on single word reading, but had more problems with continuous texts. Bishop et al. also found that NWR was unimpaired in the SLI-only group, just like the results published by Catts et al. (2005). However, the children with only SLI did show small, but significant deficits on tasks that needed phonological output and phonological memory (Bishop et al., 2009). This deficit is larger when the previous data of the children with SLI are examined. At four years of age the children with SLI score as low on NWR and PA tasks as the children with both dyslexia and SLI. This suggests that the children with SLI developed the ability to decode single words properly and therefore develop better literacy skills, which gives them the ability to score better on NWR and PA tasks later in life. However, when looking at the comprehension of children with SLI, it can be found that these children do not have good understanding of what was read (Bishop et al., 2009). So, next to oral language difficulties, children with only SLI have literacy problems as well, even though they seem to score better on the tasks than children with dyslexia and children with both dyslexia and SLI. The study conducted by Bishop et al. is as opposed to the study by Catts et al. (2005) more in line with the Multiple Deficits Model than it is with the Comorbidity Model, since they did find an overlap in the possible underlying deficit of both SLI and dyslexia.
Catts et al. (2005) and Bishop et al. (2009) disagree on the underlying deficit that might cause SLI and they disagree on how dyslexia and SLI are connected to one another. According to Catts et al., empirical findings are in support of the Comorbidity Model. Whereas Bishop et al. found their findings to be more in support of the Multiple Deficits Model. Catts et al. claim that the children that do show a phonological processing deficit are children with both dyslexia and SLI. They explain this with the less strict exclusion criteria often used to diagnose the disorders. According to Catts et al. these less strict criteria will lead to less pure forms of dyslexia and SLI. Furthermore, this will lead to
wrong results and conclusions, because the children with a pure form of SLI still might have dyslexia and vice versa. The small phonological deficit found by Catts et al. is therefore explained by them with the thin line in this overlap of dyslexia and SLI. Their exclusion criteria might still not filter out all the children with both disorders in their SLI only group. So, the children with SLI that do show deficits in PA and vSTM are children with the comorbidity of dyslexia and SLI. Catts et al. therefore propose that the PA processing is related to reading development and that the time effect either reflects on the reciprocal relation of reading and PA or indicates that PA deficits are more stable over time in children with dyslexia.
Bishop et al. (2009), however, found that children with SLI might have less persistent PA problems, which might be due to the ability to develop good literacy skills. This is supported by the data which shows that children with SLI do score well on RAN tasks. The reading ability improves the deficits in the phonology. This is why they only found small, but significant deficits on tasks that needed phonological output and phonological memory. So, RAN can be seen as a protective factor against literacy problems. Bishop et al. therefore believe that the underlying deficits that causes SLI as well as dyslexia is in phonological processing. Consequently, there is an overlap in deficits between SLI and dyslexia, namely a deficit in both PA and vSTM.
As can be seen from these two studies no conclusive answer can be given as to which model is supported best, since both studies support a different model. Hence, more research is needed with stricter exclusion criteria on the possible overlap and on the protective factors of RAN to be able to determine which model has more explanatory power.
3.2.3 SLI and ASD
There seems to be an overlap in the language profiles of children with SLI and children with ALI (Autism Spectrum Disorder with language impairment). Different studies have been conducted to determine whether language profiles indeed overlap, which might indicate that there might be a shared underlying deficit causing language problems. Two studies will be discussed here, namely the study conducted by Taylor, Maybery, Graynder & Whitehouse (2014) and the study conducted by Whitehouse et al. (2008).
Taylor et al. (2014) tried to determine whether children with ALI carry cognitive characteristics of SLI and whether children with SLI share aspects of the cognitive profiles of ASD. They tested 61 typically developing children (aged between 60 and 145 months), 19 children with SLI (aged between 63 and 145 months) and 32 children with ASD (aged between 71 and 152 months), of whom 14 had a language impairment (ALI) and 18 had normal language abilities. To test their language abilities Taylor et al. used the NWR task and the sentence repetition task (SNR) of the Developmental
be recited back by the children. The length of the sentences and non-words increased in length and difficulty.
The results of the study conducted by Taylor et al. (2014) seem to show an overlap in an underlying deficit. They found that both the children with SLI and the children with ALI showed impairments in the NWR task. The SLI also performed significantly worse on the SNR task and there was a trend for the children with ALI to perform worse than predicted too. However, this was not significant. These results could indicate an overlap in an underlying deficit causing both disorders. However, Taylor et al. were unable to compare the error patterns of the different groups due to the type of tests they used. They therefore wonder if the observed deficit in NWR in ALI is not caused by the broader developmental difficulties experienced by this group. Still the results of this study seem to be more in line with the Multiple Deficits Model.
Whitehouse et al. (2008) examined whether there is an etiological overlap between SLI and ASD. They compared the language profiles of 34 children with SLI (mean age of 11;10), 18 children with ASD without language problems (mean age of 10;11) and children with ASD and language problems (mean age of 10;8). The participants completed a battery of standardised language tests that assessed the phonological STM, verbal STM and receptive and expressive language. For example, the repetition of non-words subtest from NEPSY8 was used to test the phonological STM and the
repetition of sentences subtest from NEPSY was used to test the verbal STM.
Whitehouse et al. (2008) were able to look at the error patterns of their clinical groups. They agree that on the surface it might seem that the language profiles of children with SLI and children with ALI are the same. However, they found that children with SLI do perform significantly worse on tasks concerning the verbal STM (SNR) compared to the children with ALI and that when looking at the error patterns of NWR tasks they differ per disorder. The participants of both the SLI and the ALI group produced a similar number of errors on the two- and three-syllables words of the NWR task. However, when the non-words became longer, the words became more problematic for the children with SLI than for the children with ALI. This might indicate that children with SLI have a greater phonological STM limitation than children with ALI and that the underlying deficit that might cause ALI is not the same as that of SLI. They therefore suggest that ‘the structural language deficits – including poor non-word repetition – may arise when there is a particular conjunction of behavioural/ cognitive deficits, each of which would not disrupt non-word repetition on its own, but when they occur in combination would lead to impaired ability to do the task’ (Whitehouse et al., 2008, p.322). Hence, this study is more in line with the Comorbidity Mode, since there seems to be no overlap. The results of the study conducted by Taylor et al. (2014) are more in line with the Multiple Deficits Model and the results of the study by Whitehouse et al. (2008) match more with the
Comorbidity Model. However, both studies say that more research is needed, because their results are not completely conclusive. Taylor et al. were unable to look close enough at their data and Whitehouse et al. found that, even though their study provided preliminary support, their group size was small and more studies should be conducted with larger sample sizes.
3.2.4 Conclusion
The studies that have been discussed above show mixed results. Hence, the co-occurrence of ADHD with SLI cannot simply be explained with an overlap of a deficit in the WM. Alloway et al. (2009) and Martinussen & Tannock (2006) both found an overlap, however, both in different areas of the WM. Alloway et al. found an overlap in the verbal WM while Martinussen & Tannock found an overlap in the visuo-spatial WM and CE. In contrast, Jonsdottir et al. (2005) even found no deficit in the WM at all in either the group with ADHD and SLI or the group with ADHD only. This would suggest that a deficit in the WM is no underlying cause of ADHD or SLI at all.
The conducted studies on the co-occurrences of dyslexia and SLI are also not conclusive. Catts et al. (2005) found no overlap in the underlying deficit that might cause both dyslexia and SLI. They claimed that if there is an overlap found in a study, this is caused by the usage of less strict exclusion criteria for both SLI and dyslexia. This in turn would lead to mixed groups of SLI with dyslexia, instead of ‘pure’ forms of dyslexia and SLI and would lead to wrong results and conclusions, because the children with a pure form of SLI still might have dyslexia and vice versa. Bishop et al. (2009), however, disagree. They claim to have found an overlap in deficits in phonological processing and that these underlying deficits might cause both SLI and dyslexia. They found a deficit for both SLI and dyslexia in PA and vSTM.
The results of the studies conducted on ASD and SLI did not clearly support one of the two models either. Taylor et al. (2014) tested the vSTM with a NWR task and found that children with SLI and children with ALI scored significantly lower compared to typically developing children, which might indicate that both disorders have the same underlying deficit. However, they were unable to look at the error patterns of both groups. Whitehouse et al. (2008) were able to do so and found that the error patterns deviate from each other. This in turn indicates that there might not be an overlap in underlying deficit.
None of the studies are conclusive in their results and it is therefore not possible to determine which model, the Multiple Deficits Model or the Comorbidity Model, is supported best. More research is needed on the subject with larger group sizes, stricter criteria for the test groups and better tasks that only test what needs to be tested.
4
The Current Study
4.1 Introduction
As can be seen in the previous chapters, more research is needed on the co-occurrence of SLI with other disorders. The studies cannot give a conclusive explanation to these co-occurrences. Are there indeed overlaps in the underlying deficits or are there no overlaps at all? Another subject that is not addressed in these studies are the frequencies of these co-occurrences. However, most studies do say that the co-occurrences of these disorders with SLI happen above chance level. For example, Bishop (2010, p. 618) claims that SLI and ASD ‘co-occur at above chance level’, which according to her suggests a shared etiology.
On the basis of a Dutch data set, the aim of the current study is to answer the following questions: (1) Which disorders co-occur with SLI in the Dutch data set?
(2) What are the frequencies of these co-occurrences in the Dutch data set?
However, due to the fact that the researcher did not have the opportunity to test the participants of this study extensively, the collected data do not allow to draw conclusions on the possible underlying deficits. Therefore, the current study will only investigate the co-occurring disorders and the frequency of these co-occurrences.
4.2 Participants
The 74 participants in this study are all students of the Orion College West in Amsterdam. This a secondary school for Cluster 2 education. This means that only children attend who have a severe hearing impairment (but are not deaf) or have a severe language or communication disorder, like SLI or ASD. To attend this type of school, admission criteria need to be met. These criteria show why it is essential for these children to attend this type of school and not a mainstream school. The cluster 2 admission is valid for a limited amount of time and has to be reapplied for every two or three years. There are three types of cluster 2 admissions that can be applied for, namely:
• ASD
• Hearing impairment • SLI
For each type of application different information about the child is needed. For an ASD cluster 2 admission a psychologist needs to diagnose the child with a form of ASD, which is done with the use of various tests. For a cluster 2 admission for a child with an hearing impairment the hearing loss is tested and needs to be in a certain range. A cluster 2 admission for SLI, however, is more difficult. The
child’s language ability is tested with the use of elements of the CELF-4-NL (language structure score and language content score) and the Peabody Picture Vocabulary Test-III-NL. To fulfil the application the children need to score -1,5 standard deviation or less on both language form and content. If these criteria are not met, chances are that the application is denied. Therefore, to give the application a bigger chance of being approved, extra information about the child is included, like additional disorders the child is effected by, like ADHD, dyslexia or ADD (Attention Deficit Disorder)9.
74 children were selected for this study. All students were in their first, second or third year of secondary school. They are thought at the Dutch educational level of vmbo or havo10. They all have a
cluster 2 admission for SLI (57: male, 17: female). The mean age is 14;4 years (SD of 1;0,3 years). The children have various ethnic backgrounds (Dutch, Turkish, Portuguese) and 41 children do not have Dutch as their first language (Turkish, English, Moroccan). However, these children are not excluded from this study, since they do all qualify for a SLI cluster 2 admission and all speak Dutch. Also, the data used for this study do not include the quality of the language of the children; just the SLI cluster 2 admissions. Information about the participants can be found in Appendix I.
Permission was given by the school to use the records of their students, under the condition that the identity of the students would be kept anonymous.
4.3 Data Collection and Analysis
The data were collected from the school’s records, which contain the cluster 2 admissions and the information that is needed to reapply for a cluster 2 admission. Of all the children, only the children with a SLI cluster 2 admission were selected out of the individual records and given a number to assure anonymity. Subsequently, information on the co-occurrence of other disorders was gathered from these individual records and counted. A distinction was made between an official diagnosis of these other disorders and the clinical report of symptoms of a disorder. The frequency of the combinations of the different disorders and SLI was calculated with the use of Microsoft Office Excel 2007. Other data that were collected from the records were, age, day of birth, ethnic backgrounds and first languages.
4.4 Reliability and Validity
It has to be noted that the results might not be representative of the SLI population of the Orion
College West. It is less complicated to apply for an ASD cluster 2 admission than it is to apply for a SLI
9 ADD can be compared to ADHD-I. However, the term ADD is used for this diagnosis in the Netherlands and ADHD-I is seldom used.
cluster 2 admission. The applications for ASD are less time consuming and often more easily accepted by the commission than those for SLI. It is therefore often the case that the children with ASD are only tested for ASD and not for SLI when (re)applying for an admission. The number of children with SLI and ASD that is presented here might, consequently, be smaller than it is in reality, since there is no SLI diagnosis in the files of these children. They are therefore not included in this study.
Another problem with the validity of this study is the accuracy of the records themselves. The complete staff of the Orion College West has access to the records. Most of the time the staff members return every document after usage, however, this is unfortunately not always the case. Sometimes parts of files disappear without people noticing. Or the most recent file is taken out, leaving the researcher with an outdated file without knowing.
5
Results and Conclusion
This chapter will discuss the results of the current study. The secondary disorders that were found and the different combinations of these disorders with SLI will be discussed in section 5.1 and the frequencies and the proportions of these combinations will be given. Furthermore, section 5.2 will discussed the conclusion based on both the current study and the literature review, which have been carried out (see chapter 4 for the current study and chapter 3 for the literature review).
5.1 Co-occurrences
In the data collected, different disorders and different combinations of disorders in one individual were found. These disorders, the combinations of the disorders, the frequencies and the proportions of these combinations are shown in table 5.1.
Table 5.1 Disorders that co-occur with SLI on Orion College West. N=74 (ADHD = Attention Deficit Hyperactivity Disorder, ADD = Attention Deficit Disorder, ASD = Autism Spectrum Disorder).
Disorder(s) Frequenc y Proportion Official diagnosis Reported other disorder(s) Total SLI 26 0 26 .351 SLI Dyslexia 9 14 23 .311 SLI ASD 6 3 9 .122 SLI ADHD 3 3 6 .081
SLI Dyslexia ADHD 2 2 4 .054
SLI Dyslexia ADD 2 1 3 .041
SLI Dyslexia ASD 1 1 2 .027
SLI Dyslexia ASD ADHD 1 0 1 .014
Total 50 24 74 1.000
Secondary disorders that were found along with SLI were: ASD, ADHD, ADD and dyslexia. In table 5.1 the group of students with only SLI is the largest, followed by the group of students with SLI and dyslexia. The frequency and the proportion get lower when more additional disorders are found. Some of the students have two or more additional disorders in combination with SLI.
5.2 Conclusion
This thesis tried to answer the following questions: (1) Which disorders co-occur with SLI?
(3) Can these co-occurrences be explained by a common underlying deficit?
In order to answer these questions a literature review and a small desk study were conducted. Based on this it was found that more research is needed on the subject of co-occurring disorders and their underlying deficits.
The disorders that were mainly found to co-occur with SLI in the literature review and in the desk study were ADHD, dyslexia and ASD. Other disorders co-occur with SLI as well, like motor problems (Bishop, 2002), but little research is done on these co-occurrences. This is why these three disorders were chosen to elaborate on in the theoretical framework of this thesis. The desk study confirmed these three disorders to co-occur with SLI. What was also found in the desk study was that multiple disorders can co-occur in one individual. For instance, up to four co-occurring disorders were found in one individual (student 11, table 1, Appendix I).
The studies used as a framework for the literature review did not elaborate on the frequencies of the occurrences of SLI with other disorders. However, most studies indicate that the co-occurrences of these disorders with SLI are statistically significant. For example, Bishop (2010, p. 618) claims that SLI and ASD ‘co-occur at above chance level’, which according to her suggests a shared etiology. The desk study was able to elaborate more on the subject. The results were shown in table 5.1. This table shows that the group of children with both dyslexia and SLI is the largest, followed by the group of children with ASD and SLI. The frequency and the proportion get lower when more additional disorders are found.
The group size of the children with both dyslexia and SLI might be explained by the similarities between the two disorders, since both disorders affect language (see 3.2.2). Some researchers state that 53% of the children with a reading disability classify for SLI and 55% of children with SLI classify for a reading disability (McArthur et al., 2000). However other researchers argue that these numbers are far too high and are caused by the use of lenient exclusion criteria for both disorders (Catts et al., 2005). It is not always clear if a child has SLI or dyslexia or both and that therefore children are more frequently tested for dyslexia than for any of the other co-occurring disorders that were found. However, the results concerning the children with SLI and ASD might not be representative of the SLI population on the Orion College West. The criteria that were used to select the children included that the children needed to have a SLI cluster 2 admission. However, it is less complicated to apply for an ASD cluster 2 admission than it is to apply for a SLI cluster 2 admission. It is therefore often the case that children with ASD are only tested for ASD and not for SLI when (re)applying for an admission. This means that for this study the number of children with both SLI and ASD might be smaller than it is the case in reality, since these children are not selected due to the lack of a SLI cluster 2 admission in their files.
What also needs to be kept in mind, when interpreting the data, is the validity of the current study. The records are accessible by all the members of the staff of the Orion College West. It is therefore possible that some records were taken out by a staff member during the study or that some parts of the files were misplaced or never returned.
Unfortunately, nor the current study nor the literature review were able to answer the last question:
(3) Can these co-occurrences be explained by a common underlying deficit?
Unfortunately, it turned out to be impossible to answer the last question with the data of the current study, due to the fact that the participants would need to be retested extensively, which, for several reasons, was not feasible. The studies from the literature were also unable to give a conclusive answer to the third question. This makes it clear that more research is needed to investigate the precise underlying deficit of each disorder and on the possible commonalities in these underlying deficits. Are the different disorders indeed related through an underlying deficit (the Multiple Deficits Model) or are the disorders complete separate entities (the Comorbidity Model)?
Even though no answer can be given to the third question, the answers to the other two questions can help us in the future with determining a diagnosis. The high frequency of certain combinations in disorders might indicate that the chance of these disorders to co-occur is fairly high. For instance, if a child has SLI, chances are that it might also have dyslexia or ASD or even both. Of course, this is not always the case, but we should be more aware of the possibility. This in turn might help with an earlier diagnosis, so more suitable help can be offered to the child at a younger age.
6
Discussion and Future Directions
Both the current study and the literature review show that more research is needed on the subject of the co-occurrence of SLI with other disorders. Most researchers agree that the disorders mentioned can co-occur with each other in one individual. However, on the subject of why and how these disorders co-occur, no consensus is reached.
The literature review gave insight in different studies, all with different outcomes. Some studies were more in line with the Multiple Deficits Model (Alloway et al., 2009, Martinussen & Tannock, 2006, Bishop et al., 2009 and Taylor et al., 2014), while others were more in line with the Comorbidity Model (Jonsdottir et al., 2005, Catts et al., 2005 and Whitehouse et al., 2008). Moreover, there even was no consensus on the underlying deficit that might cause the disorder. For example, Jonsdottir et al. (2005) found no deficit in the WM of the children with ADHD, while this is often considered to be one of the underlying deficits of ADHD.
The results of the current study were unable to help clarify the different results found in the literature review, since the results of the current study could only depict the frequency of the co-occurrences of the different disorders with SLI. Therefore a small research proposal was written to show which other data would be needed to help answer the third question: Can these co-occurrences be explained by a common underlying deficit?
6.1 Research Proposal
In order to answer the third question, other data are needed from, for example, the students of
Orion College West. The children would need to be tested extensively and consistently, especially
since some of the children have been diagnosed with more than two disorders.
6.1.1 Participants
Four groups of participants will be used to conduct this study. The first group of participants will be the children of the Orion College West with a SLI cluster 2 admission and one or more additional diagnosed disorder(s) (ADHD, dyslexia and/or ASD). The children that do not have an official diagnosis are excluded.
The second group of participants will also be children of the Orion College West. This group will exist of children with only SLI. The children therefore will have a SLI cluster 2 admission. When there is a suspicion of another co-occurring disorder the child is excluded from the study.
The third group of children will be aged matched children from (mainstream) secondary schools with an official diagnosis of one of the following disorders:
• ADHD • Dyslexia
• ASD with language impairment (ALI)
The participants of this group may not have any other disorder. They therefore have to be tested on, for example, their language abilities to exclude SLI. If any of these participants show language problems typical of SLI, they are excluded.
The last group will exist of typically developing children matched for age. These participants are excluded, when there is a suspicion of SLI, ADHD, dyslexia or ASD.
Other exclusion criteria would be low intelligence (nonverbal, verbal and total IQ below 85), hearing loss or any other problems that could cause language problems. All participants should have the same first language, since their language abilities are tested in this study.
6.1.2 Design
Different tests need to be used for the different disorders. The children who are diagnosed with dyslexia will be tested on their phonological processing, since that is the area were possible overlap in the underlying deficit is found between dyslexia and SLI. The children who are diagnosed with ADHD will be tested on their working memory. Although the studies discussed in section 3.2.3 have not found a conclusive answer to whether or not the underlying deficit is found in this area, this study could be a good opportunity to carry out more research on this topic. The children with ASD (ALI) will be tested on verbal short-term memory. All tests need to be done in the typically developing group and the group of children with only SLI.
The tests concerning ADHD will be based on the tests conducted by Martinussen & Tannock (2006). The domains that will be tested of the WM are the verbal WM, visuo-spatial WM and the CE. The last domain is subdivided into a verbal and a spatial component. The verbal WM and CE will be tested with the Digit Forward and the Digit Backward subtests from the WISC-III. The spatial WM and CE will be tested with two tasks which require the participants to remember and reproduce a sequence of spatial locations in either the same or reverse order (Martinussen & Tannock, 2006). The VS will be tested with the Finger Windows task from the Wide Range Assessment of Memory and
Learning (WRAML; Adams and Sheslow, 1990). In this task the participant will be asked to exactly
reproduce a demonstrated spatial sequence on a card. These sequences will increase in length. This task will be done in reverse to tap into the spatial CE.
The test concerning dyslexia will be based on the tests conducted by Vandewalle et al. (2012). With these tests PA, vSTM and lexical access is tested. PA will be tested with a phoneme deletion task and a spoonerism task. The phoneme deletion task will exist of nonwords of which a certain phoneme needs to be deleted. This will lead either to an existing word or another nonword. The
