Editorial: Seeking a new characterisation of learning disorders

Abstract

For many years, educationists, dissatisfied with the medical model, have disputed the use of diagnostic categories. Rather than ‘dyslexia’, ‘ADHD’ or ‘Specific language impairment (SLI)’, terms such as literacy difficulties, problems of attention, and speech, language and communication needs (SLCN) abound in our schools. A similar tension between categorical and dimensional disorders has been at the forefront in the development of DSM-5, and is a recurrent theme in several articles in this issue. Together these articles highlight the need for a broader perspective that recognizes continuities and co-morbidities between children’s learning disorders. Van Bergen and colleagues (Van Bergen et al., 2012) studied children at family risk (FR) of dyslexia to characterise the differences between children from ‘high-risk’ families who were or were not classified as ‘dyslexic’ at 8 years of age. Thirty-one per cent of the FR group were considered dyslexic after two years of reading instruction (defined as scoring below the 10th centile in word reading fluency); as expected, these children performed significantly below the level of controls from families with no family history of dyslexia on tests of phonological awareness, rapid naming (RAN) and spelling. Importantly, however, the FR group who did not fulfil the criteria for dyslexia were also impaired but to a lesser degree on tests of reading fluency and phonological awareness relative to controls. This finding highlights the fact that dyslexia is not an ‘all or none’ condition. Rather, dyslexia represents the low end on a continuum of reading with milder or ‘sub-clinical’ forms of reading disorder sharing the same core characteristics. Preliminary analyses using data from affected parents suggested that one of the factors determining where an individual lies on the liability continuum is parental reading skills; because data are available from only one parent in each family, further evidence is required to test this hypothesis. A similar picture emerges from the behaviour-genetic study of childhood attention-deficit hyperactivity disorder (ADHD) by Larsson et al. (2012). This study used quantitative measures of ADHD symptoms (derived from telephone interviews with parents) to investigate the etiology of ADHD in a large epidemiological sample of some 16,000 twins aged 9 and 12 years. Using both extreme groups analysis and analysis of individual differences across the full range of the distribution, the article reports that there was no difference in etiology between ADHD defined by standard diagnostic criteria and sub-clinical variants of the disorder seen in children whose symptom scores were below the clinical threshold. Thus, ADHD is best viewed as the extreme of a dimension of ADHD symptoms with the same genetic and environmental factors operating to cause the full range of symptoms of inattention, hyperactivity and impulsivity as well as milder levels of symptomatology. Henry et al. (2012) consider specific language impairment (SLI), a neurodevelopmental disorder which significantly impacts learning in a number of domains and is frequently co-morbid with developmental coordination disorder. The focus of their article is on one aspect of the cognitive profile of these children, executive functioning (EF). The clinical sample for the study was large by current standards, though it should be noted that the age range was wide. There were three groups of participants: a group of children with SLI who had formal clinical diagnoses and also fulfilled research criteria (with nonverbal skills within the normal range); a comparison group of typically developing children, and a further group described as ‘low language functioning’ (LLF) who had milder language difficulties, in some cases alongside low nonverbal IQ. The assessment battery was comprehensive and included verbal and non-verbal versions of EF tasks tapping five constructs: executive loaded working memory, fluency, inhibition, planning and switching. The SLI group performed worse than controls on 6 of the 10 EF measures, showing impairments in verbal and non-verbal working memory, verbal and non-verbal fluency, non-verbal inhibition and non-verbal planning. There was also a trend for the SLI group to score poorly in verbal inhibition, verbal planning and non-verbal switching, leaving only verbal switching (a test of relatively low reliability) unaffected. Importantly, for all tasks except for verbal fluency, the SLI and LLF groups performed similarly. One interpretation of this finding is that some degree of language difficulty is sufficient to depress performance on EF tasks, for example, by suppressing the use of verbal mediation. The authors tested this hypothesis by controlling for age as well as verbal and non-verbal IQ; significant group differences remained for verbal and non-verbal WM, verbal fluency, non-verbal inhibition and non-verbal planning, and again there was no difference between the SLI and LLF groups. Although the term ‘specific’ is used in the classification of children with significant language delay who have normal nonverbal IQ, the study of Henry et al., bolsters evidence that the problems of these children are not specific to language. An intriguing hypothesis is that the broad range of EF deficits observed in this study go some way toward explaining their frequently co-occurring motor impairments and pragmatic difficulties. However, evidence of causal relationships is lacking and longitudinal studies bearing on this issue are badly needed. The findings are consistent with a dimensional view in that there was no difference in performance between children with SLI, and those with less specific language problems. Here, as with dyslexia and ADHD, the cut-off between impaired and typical functioning is difficult to operationalise. DSM-5 appears to circumvent this issue by proposing the use of the overarching term ‘language impairment’ until (or if) a more specific subtype is diagnosed. Finally Van Hulle and colleagues (Van Hulle et al., 2012) alert us to the importance of recognizing sensory over-responsivity (SOR) as a condition that can lead to a range of difficulties affecting school adjustment and educational achievement. Not only does this disorder occur independently of common childhood psychiatric diagnoses, but it also commonly co-occurs with other disorders – in this study, more than 60% of 7-year-old twins with a clinical ‘diagnosis’ also screened positive for SOR. Moreover, the heritability of SOR is high and it is largely genetic factors that account for its co-morbidity with other disorders. Bringing these findings together, it is impossible to deny the long-held view of educationalists that disorders are dimensional. But equally, to contend that the disorders which affect children’s learning are merely extremes on continuous dimensions is too simplistic. Cognitive impairments of all types affect children’s learning and adjustment in school, and it is important for clinicians to know that individual differences in the behavioural manifestations of disorders are the outcome of multiple risks which act probabilistically to confer liability for a given disorder. Hence, whether or not a child reaches threshold for ‘Disorder A’ (say dyslexia) may depend not only upon their phonological skills (an established causal factor) but also where they fall on a dimension of Disorder B (say an endophenotype of language impairment), in interaction with where they fall on a Dimension C (say an endophenotype underpinning ADHD). In the face of such complexity, it is hard to deny that diagnostic labels offer clarity to the field. Such labels offer a means of organizing together the core characteristics of a dimensional disorder. In the case of disorders that affect learning, these labels also communicate that a child has additional needs. In fact, when labels are unfamiliar or used inconsistently, the timely identification of a child’s educational difficulties will be compromised and the disability may remain hidden, as is the case for SOR, and arguably also SLI, a disorder with considerable impact on educational achievement which is still poorly recognized (Bishop, 2010).