* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download Chapter 2 - Kenniscentrum Kinder
Narcissistic personality disorder wikipedia , lookup
Diagnostic and Statistical Manual of Mental Disorders wikipedia , lookup
Classification of mental disorders wikipedia , lookup
History of psychiatry wikipedia , lookup
Generalized anxiety disorder wikipedia , lookup
Dementia with Lewy bodies wikipedia , lookup
Conduct disorder wikipedia , lookup
Pyotr Gannushkin wikipedia , lookup
Conversion disorder wikipedia , lookup
Dissociative identity disorder wikipedia , lookup
Abnormal psychology wikipedia , lookup
Spectrum disorder wikipedia , lookup
Child psychopathology wikipedia , lookup
Autism therapies wikipedia , lookup
Asperger syndrome wikipedia , lookup
Controversy surrounding psychiatry wikipedia , lookup
Sluggish cognitive tempo wikipedia , lookup
Autism spectrum wikipedia , lookup
Attention deficit hyperactivity disorder wikipedia , lookup
Attention deficit hyperactivity disorder controversies wikipedia , lookup
Behavioral and Cognitive Profiling in Autism Spectrum Disorder and Attention-Deficit/Hyperactivity Disorder Jolanda Maria Johanna van der Meer Behavioral and Cognitive Profiling in Autism Spectrum Disorder and Attention-Deficit/Hyperactivity Disorder Een wetenschappelijke proeve op het gebied van de Medische Wetenschappen Academisch Proefschrift Cover design Hartebeest Cover photo (back side) Charelle Fotografie ter verkrijging van de graad van doctor Layout and print aan de Radboud Universiteit Nijmegen Proefschriftmaken.nl – Uitgeverij BOXPress op gezag van de rector magnificus prof. mr. S.C.J.J. Kortmann, volgens besluit van het college van decanen in het openbaar te verdedigen op woensdag 3 september 2014 Support om 14.30 uur precies This PhD project was supported by Karakter Child and Adolescent Psychiatry, and the Netherlands Organisation for Scientific Research (NWO) by grants assigned door to Buitelaar (05613015) and Rommelse (91610024). Jolanda Maria Johanna van der Meer Copyright © J.M.J. van der Meer, 2014. All rights reserved. No part of this thesis may be reproduced, stored in a retrieval system or transmitted in any form or by any means without prior written permission of the author. Geboren op 6 september 1983 te Enschede. Promotor Prof. dr. J.K. Buitelaar Copromotoren Dr. N.N.J. Lambregts-Rommelse Dr. C.A. Hartman (Rijksuniversiteit Groningen) Manuscriptcommissie Prof. dr. M. Willemsen (voorzitter) Prof. dr. H. Bekkering Prof. dr. H. Roeyers (Universiteit Gent, België) Now nature never deals in black or white. It is always some shade of grey. She never draws a line without smudging it. Winston S. Churchill Winston And Clementine: The Personal Letters Of The Churchills Table of Contents Chapter 1 General introduction, aims and outline of the thesis 11 Chapter 2 Are high and low extremes of ASD and ADHD trait continua pathological? A population-based study using the AQ and SWAN rating scales 35 Chapter 3 Are autism spectrum disorder and attention-deficit/hyperactivity disorder different manifestations of one overarching disorder? Cognitive and symptom evidence from a clinic and population-based sample 61 Chapter 4 How ‘core’ are motor timing difficulties in ADHD? A latent class comparison of pure and comorbid ADHD classes 91 Chapter 5 Homogeneous combinations of ASD-ADHD traits and their cognitive and behavioral correlates in a population-based sample 111 Chapter 6 Using cognitive profiles to examine the relationship between ASD and ADHD 133 7 Chapter 7 A randomized, double-blind comparison of atomoxetine and placebo on response inhibition and interference control in children and adolescents with autism spectrum disorder and comorbid attentiondeficit/hyperactivity disorder symptoms 161 Chapter 8 General discussion, summary, discussion, key findings, limitations, future directions and clinical implications 183 Chapter 9 Samenvatting in het Nederlands (Summary in Dutch) 211 Chapter 10 References223 Chapter 11 Dankwoord (Acknowledgements in Dutch) 267 Chapter 12 About the author 8 275 9 General introduction 10 This thesis focuses on the shared and unique behavioral and cognitive profiles of Autism Spectrum Disorder (ASD) and Attention-Deficit/Hyperactivity Disorder (ADHD). Studying ASD and ADHD together may provide the most optimal strategy in examining both shared and unique substrates, ultimately translating into differential prognoses and susceptibility towards treatment. The current research approach steps away from the Diagnostic and Statistical Manual of Mental Disorders (DSM) defined heterogeneous group comparisons, and acknowledges the continuously distributed nature of the ASD and ADHD trait within the population, as well as the etiological and symptomatic heterogeneity within both disorders. In this general introduction, the co-occurrence, etiology and treatment of ASD, ADHD and related cognitive profiles are discussed. Then, the research approach used to reduce heterogeneity on both the behavioral and cognitive level is described. Finally, the outline of the chapters is provided. ASD and ADHD With prevalence rates of about 1% for ASD and 5% for ADHD, these disorders are among the most commonly diagnosed psychiatric developmental disorders in children and adolescents (Baird et al., 2006; Polanczyk, de Lima, Horta, Biederman & Rohde, 2007). ASD is characterized by impaired social interaction skills and verbal and nonverbal communication, as well as restricted and repetitive behavior and interests, while ADHD is characterized by severe inattention, hyperactivity and impulsivity (American Psychiatric Association, 2013). Symptom presentations of both disorders are rather heterogeneous, as described in the DSM (American Psychiatric Association, 2013), see Box 1.1. The DSM-IV and previous psychiatric classification schemes prevented a diagnosis of ADHD in the context of ASD. This prohibition was based on the assumption that ASD is an overarching disorder that mimics or even causes symptoms of ADHD. As a consequence, patients were diagnosed with either ASD or ADHD, disregarding possible co-occurring symptoms. The heterogeneous symptom presentation on the one hand and the prohibited comorbid diagnosis of ASD and ADHD on the other hand may explain 13 Chapter 1 the not-to-be-ignored proportion of children that have been alternatively given a diagnosis of one or the other disorder throughout development (Fein, Dixon, Paul & Levin, 2005). In the current DSM-5, a comorbid diagnosis of ASD and ADHD can be made (American Psychiatric Association, 2013). This step forward will boost research on the shared and specific underlying mechanisms related to ASD and ADHD, and can inform us on the association between both disorders. Even though the diagnostic criteria appear to show little overlap, symptoms of ASD and ADHD may be entangled. For example, inattention can easily be mistaken for social inattention, and stereotyped behaviors (such as body rocking and hand flapping) may be mistaken for hyperactivity. Although such entangled symptoms may result in inflated ASD-ADHD comorbidity rates, factor analyses found no overlapping diagnostic criteria, which supports the Box 1.1 DSM-5 diagnostic criteria for ASD and ADHD Autism Spectrum Disorder (ASD) Diagnostic Criteria A. Persistent deficits in social communication and social interaction across multiple contexts, as manifested by the following, currently or by history (examples are illustrative, not exhaustive): 1. Deficits in social-emotional reciprocity, ranging, for example, from abnormal social approach and failure of normal back-and-forth conversation; to reduced sharing of interests, emotions, or affect; to failure to initiate or respond to social interactions. 2. Deficits in nonverbal communicative behaviors used for social interaction, ranging, for example, from poorly integrated verbal and nonverbal communication; to abnormalities in eye contact and body language or deficits in understanding and use of gestures; to a total lack of facial expressions and nonverbal communication. 3.Deficits in developing, maintaining, and understanding relationships, ranging, for example, from difficulties adjusting behavior to suit various social contexts; to difficulties in sharing imaginative play or in making friends; to absence of interest in peers. independence of ADHD and ASD diagnostic criteria (Ghanizadeh, 2010; Martin, Specify current severity level 1 / 2 / 3, see below. Hamshere, O’Donovan, Rutter & Thapar, 2014; see for review Rommelse, Franke, B. Restricted, repetitive patterns of behavior, interests, or activities, as manifested by at least two of the following, currently or by history (examples are illustrative, not exhaustive): Geurts, Hartman & Buitelaar, 2010). Hence, the co-occurrence of ASD and ADHD is unlikely to be largely due to mistaken symptom interpretations. In clinic based 1. Stereotyped or repetitive motor movements, use of objects, or speech (e.g., simple motor stereotypies, lining up toys or flipping objects, echolalia, idiosyncratic phrases). samples, the majority of comorbidity estimates reported for ADHD in ASD fall 2. Insistence on sameness, inflexible adherence to routines, or ritualized patterns of verbal or nonverbal behavior (e.g., extreme distress at small changes, difficulties with transitions, rigid thinking patterns, greeting rituals, need to take same route or eat same food every day). within the range of 30% to 80%, whereas the presence of ASD is estimated in 20% to 50% of the patients with ADHD (e.g. Ames & White, 2011; Leyfer et al., 2006; for review see Rommelse et al., 2010; Ronald, Simonoff, Kuntsi, Asherson & Plomin, 2008). The importance of comorbidity in taxonomic questions forms the basis of critical hypotheses in both research and clinical practice, as was already 3. Highly restricted, fixated interests that are abnormal in intensity or focus (e.g, strong attachment to or preoccupation with unusual objects, excessively circumscribed or perseverative interest). 4. Hyper- or hyporeactivity to sensory input or unusual interests in sensory aspects of the environment (e.g., apparent indifference to pain/temperature, adverse response to specific sounds or textures, excessive smelling or touching of objects, visual fascination with light or movement). described decades ago (Caron & Rutter, 1991; Neale & Kendler, 1995). Provided Specify current severity level 1 / 2 / 3, see below. that comorbidity is not due to artifacts such as chance, sampling bias, population C.Symptoms must be present in the early developmental period (but may not become fully manifest until social demands exceed limited capacities, or may be masked by learned strategies in later life). stratification or symptom overlap, perhaps the most fundamental issues are at the nosological level: Are the two disorders distinct, or do they reflect an arbitrary division of a single syndrome (Neale & Kendler, 1995). True comorbidity may be due to either shared or related risk factors, to a comorbid pattern constituting a meaningful syndrome, or to one disorder creating an increased risk for the 14 D. Symptoms cause clinically significant impairment in social, occupational, or other important areas of current functioning. E. These disturbances are not better explained by intellectual disability (intellectual developmental disorder) or global developmental delay. Intellectual disability and autism spectrum disorder frequently co-occur; to make comorbid diagnoses of autism spectrum disorder and intellectual disability, social communication should be below that expected for general developmental level. 15 Chapter 1 Box 1.1 Continued. DSM-5 diagnostic criteria for ASD and ADHD Box 1.1 Continued. DSM-5 diagnostic criteria for ASD and ADHD Severity is based on social communication impairments (A) and restricted, repetitive patterns of behavior (B): Attention-Deficit/Hyperactivity Disorder (ADHD) Severity level Social communication Restricted, repetitive behaviors Level 3 “Requiring very substantial support” Severe deficits in verbal and nonverbal social communication skills cause severe impairments in functioning, very limited initiation of social interactions, and minimal response to social overtures from others. For example, a person with few words of intelligible speech who rarely initiates interaction and, when he or she does, makes unusual approaches to meet needs only and responds to only very direct social approaches Inflexibility of behavior, extreme difficulty coping with change, or other restricted/repetitive behaviors markedly interfere with functioning in all contexts. Great distress/difficulty changing focus or action. Level 2 “Requiring substantial support” Marked deficits in verbal and nonverbal social communication skills; social impairments apparent even with supports in place; limited initiation of social interactions; and reduced orabnormal responses to social overtures from others. For example, a person who speaks simple sentences, whose interaction is limitedto narrow special interests, and who has markedly odd nonverbal communication. Inflexibility of behavior, difficulty coping with change, or other restricted/repetitive behaviors appear frequently enough to be obvious to the casual observer and interfere with functioning ina variety of contexts. Distress and/or difficulty changing focus or action. Without supports in place, deficits in social communication cause noticeable impairments. Difficulty initiating social interactions, and clear examples of atypical or unsuccessful response to social overtures of others. May appear to have decreased interest in social interactions. For example, a person who is able to speak in full sentences and engages in communication but whose conversation with others fails, and whose attempts to make friends are odd and typically unsuccessful. Inflexibility of behavior causes significant interference with functioning in one or more contexts. Difficulty switching between activities. Problems of organization and planning hamper independence. Level 1 “Requiring support” 16 A. Either (1) or (2): (1)Six or more symptoms of inattention for children up to age 16; symptoms of inattention have been present for at least 6 months, and they are inappropriate for the developmental level: Inattention a. Often fails to give close attention to details or makes careless mistakes in schoolwork, at work, or with other activities. b. Often has trouble holding attention on tasks or play activities. c. Often does not seem to listen when spoken to directly. d. Often does not follow through on instructions and fails to finish schoolwork, chores, or duties in the workplace (e.g., loses focus, side-tracked). e. Often has trouble organizing tasks and activities. f. Often avoids, dislikes, or is reluctant to do tasks that require mental effort over a long period of time (such as schoolwork or homework). g. Often loses things necessary for tasks and activities (e.g. school materials, pencils, books, tools, wallets, keys, paperwork, eyeglasses, mobile telephones). h. Is often easily distracted. i. Is often forgetful in daily activities. (2)Six or more symptoms of hyperactivity-impulsivity for children up to age 16; symptoms of hyperactivity-impulsivity have been present for at least 6 months to an extent that is disruptive and inappropriate for the developmental level: Hyperactivity a. Often fidgets with or taps hands or feet, or squirms in seat. b. Often leaves seat in situations when remaining seated is expected. c. Often runs about or climbs in situations where it is not appropriate (adolescents or adults may be limited to feeling restless). d. Often unable to play or take part in leisure activities quietly. e. Is often “on the go” acting as if “driven by a motor”. f. Often talks excessively. Impulsivity g. Often blurts out an answer before a question has been completed. h. Often has trouble waiting his/her turn. i. Often interrupts or intrudes on others (e.g., butts into conversations or games). In addition, the following conditions must be met: B. Several inattentive or hyperactive-impulsive symptoms were present before age 12 years. C. Several symptoms are present in two or more settings (e.g., at home, school or work; with friends or relatives; in other activities). D. There is clear evidence that the symptoms interfere with, or reduce the quality of, social, school, or work functioning. 17 as alternate expressions of a single underlying dimension of liability, or as symptomatic phenocopies. 18 Note. These models present ASD and ADHD as a) true comorbidity, b) comorbidity reflecting three independent disorders, c/d) alternate expressions of a single underlying dimension of liability, e) symptomatic phenocopies. Not presented is a possible model comorbidity due to reciprocal causation, in which ASDincreases the risk for ADHD or vice versa (‘domino effect’). as reflecting three independent disorders (ASD, ADHD, or ASD plus ADHD), Behavioral symptoms Figure 1.1. Roughly, these models present ASD and ADHD as truly comorbid, Cognitive impairment A slightly adapted version of these models for ASD and ADHD is presented in Abnormal brain conditions & Roessner (2007) also described several models of non-artifactual comorbidity. Genetic and environmental factors other (Caron & Rutter, 1991). More recently, Banaschewski, Neale, Rothenberger Comorbid condition as a) true comorbidity b) independent disorders Based on the types of symptoms, three kinds (presentations) of ADHD can occur: - Combined Presentation: if enough symptoms of both criteria inattention and hyperactivity-impulsivity were present for the past 6 months. - Predominantly Inattentive Presentation: if enough symptoms of inattention, but not hyperactivity-impulsivity, were present for the past six months - Predominantly Hyperactive-Impulsive Presentation: if enough symptoms of hyperactivity-impulsivity but not inattention were present for the past six months. Figure 1.1 Models of comorbidity in ASD and ADHD, inspired by Banaschewski et al. (2007) E. The symptoms do not happen only during the course of schizophrenia or another psychotic disorder. The symptoms are not better explained by another mental disorder (e.g. Mood Disorder, Anxiety Disorder, Dissociative Disorder, or a Personality Disorder). c) common underlying d) common underlying e) symptomatic phenocopies dimension of liability dimension of liability Chapter 1 19 Chapter 1 An increasing number of studies documented on various patterns of Devincent & Drabick, 2008; Mulligan et al., 2009), anxiety, depression and tic association between ASD and ADHD. It has for instance been hypothesized that disorders (e.g. Simonoff et al., 2008), sleep disorders (for review see Ming & ASD and ADHD may both be manifestations of the same overarching disorder Walters, 2009), and motor coordination deficits (e.g. Reiersen, Constantino & (Rommelse, Geurts, Franke, Buitelaar & Hartman 2011; Taurines et al., 2012). Todd, 2008). Children with more comorbid difficulties reportedly experience more According to this hypothesis, ADHD can be seen as a mild, less impaired subtype difficulty in daily life as compared to those with one disorder (e.g. Bauermeister within the more severe ASD spectrum, see also Figure 1.2 that was published by et al., 2007; Goldstein & Schwebach, 2004; Kotte et al., 2013). It is therefore Rommelse et al. (2011). This model suggests that the genetic substrates of ‘pure’ important to assess ASD and ADHD symptoms in parallel, while co-occurring ADHD versus ASD with comorbid ADHD are either largely similar (the difference psychiatric symptoms should be accounted for, as has been done in this thesis. in severity between the two primarily stems from environmental factors), or that the genes involved in ADHD are a subset of those involved in ASD (with difference Genetics of ASD and ADHD in severity explained by an additional biological risk for ASD, and their interactions ASD and ADHD are both genetically determined disorders, with moderate to high with environmental factors). If true, this model may have implications for future heritability estimates (e.g. Franke et al., 2012; Hallmayer et al., 2011). For ADHD, diagnostic procedures and treatment, as treatment successful in improving an estimated 76% of the phenotype is explained by its genetics, in ASD these ADHD symptomatology may also prove beneficial in ASD (Davis & Kollins, 2012; estimates vary between 60% and 90% (e.g. Faraone, Perlis et al., 2005; Freitag, Sikora, Vora, Coury & Rosenberg, 2012). A variant of this hypothesis is that ASD 2007). Several studies indicated that the frequent comorbidity of both disorders is and ADHD do not constitute different expressions of one overarching disorder, likely to be related to a modest to substantial overlap in genetic and environmental but rather reflect the presence of two distinct disorders with a common etiology risk factors (Lichtenstein, Carlstrom, Rastam, Gillberg & Anckarsater, 2010; (distinct disorders hypothesis) (Grzadzinski et al., 2011) Mulligan et al., 2009; Nijmeijer et al., 2010; Rommelse et al., 2010; Ronald et al., 2008; 2010; St. Pourcain et al., 2011; Stam, Schothorst, Vorstman & Staal, 2009; Figure 1.2 Overarching disorder hypothesis, inspired by Rommelse et al. (2011) ADHD ASD ASD and ADHD may be different manifestations of the same overarching disorder. The manifestations may range from ADHD with few if any social problems, through ADHD with greater levels of social and communicative problems, to ASD as the most severe subtype characterized by additional more severe social and communicative problems mostly combined with various levels of ADHD symptoms. Taylor et al., 2013). However, a recent study on the largest data set currently available from psychiatric genome-wide association studies (GWAS) did not use quantitative genetic methods but assessed SNP-heritability and was unable to find shared genes for ASD and ADHD (Cross-Disorder Group of the Psychiatric Genomics, 2013). Authors suggested that the found absence of genetic overlap may be due to the presence of ASD-affected children in previous ADHD-cohort studies and vice versa, which may have inflated the expected shared genetic risk factors for ASD and ADHD. This seems too early a conclusion, since shared gene- In addition to the comorbidity of ASD and ADHD, both disorders are environment interactions were not included in this study, rare genetic variants also frequently associated with other neurodevelopmental disorders such as rather than common SNPs may contribute to the overlap in genetic factor (for oppositional defiant disorder (ODD) / conduct disorder (CD) (e.g. Gadow, review see McCarthy et al., 2008), and still little is known on the specific genes 20 21 Chapter 1 involved in ASD and ADHD, let alone those involved in their overlap. At any rate, Multiple ASD and ADHD riskgenes research on the overlap in genetic and environmental risk factors for ASD and ADHD may benefit from alternative classifications and group comparisons. Such a classification should ideally move beyond descriptive syndromes, an approach applied in the chapters 3 to 6 of this thesis. Intermediate Intermediate Phenotype A Phenotype B Cognition in ASD and ADHD ASD and/or ADHD Cognitive measures are used frequently in the assessment of ASD and ADHD. behavioral symptom These measures are also referred to as intermediate phenotypes, useful indicators Figure 1.3 Etiological model of ASD and ADHD, inspired by Franke and colleagues (2009) Schematic representation of the intermediate phenotype concept in psychiatric genetics. Many genes are involved in causing an ASD/ADHD behavioral symptom, while a reduced number of genes is involved in related intermediate phenotypes such as cognitive functioning. in detecting etiologically more homogeneous subgroups of patients (Gottesman & Gould, 2003). Intermediate phenotypes form a causal link between genes and behavioral symptoms, more closely linked to the genes in action in ASD and of how neurocognitive impairments form a link between susceptibility genes ADHD, and more objectively measured than behavior (Kendler & Neale, 2010). and ASD and ADHD symptoms such as hyperactivity, impulsivity, rigidity and Characteristic of intermediate phenotypes is that they are heritable, associated impaired social interaction skills. It should be noted that this model is a simplified with the disorder, state independent and present in non-affected family members representation of reality. Most importantly, the model does not account for the of patients (Walters & Owen, 2007). A schematic representation of the intermediate influence of potential environmental and genetic risk or protective factors (Jaffee phenotype concept was published by Franke, Neale & Faraone (2009), an & Price, 2007), nor for the influence of the aforementioned frequently associated adapted version is presented in Figure 1.3. difficulties. Although simplified, this concept provides a comprehensive framework The intermediate phenotype concept provides an interpretable model for exploring the relationship between the behavioral and cognitive profiles seen in ASD and ADHD, as was done in the chapters 3 to 6 of this thesis. Cognitive impairments that characterize ASD are reported most frequently in the cognitive domains of emotion recognition (also referred to as social cognition), central coherence, cognitive flexibility and other aspects of executive functioning (e.g. Booth, Charlton, Hughes, Happé, 2003; Booth & Happé, 2010; Corbett, Constantine, Hendren, Rocke & Ozonoff, 2009; Pellicano, Maybery, Durkin & Maley, 2006). In contrast, ADHD is most frequently related to cognitive difficulties in variability, inhibitory control, motor speed and (visual) attention and working memory (for review see Doyle, 2006; Halperin, Trampush, Miller, Marks & Newcorn, 2008; Hervey et al., 2006; for review see Kasper, Alderson, & Hudec, 2012; for review see Kofler et al., 2013). Furthermore, attention problems, 22 23 Chapter 1 language delay and pragmatic language problems, sensory overresponsivity, Tavare & Gringras, 2006; Stigler, Desmond, Posey, Wiegand & McDougle, 2004). motor problems and deficits in executive functioning are frequently disclosed in Overall findings suggest that pharmacological treatment for ADHD symptoms is both ASD and ADHD (e.g. Geurts, Verté, Oosterlaan, Roeyers & Sergeant, 2004; indeed effective in reducing ADHD symptoms in patients with ASD and ADHD. Leonard, Milich & Lorch, 2011; Mulligan et al., 2009; Nydén et al., 2010; for review However, described benefits are smaller and adverse effects tend to be more see Rommelse et al., 2011). This growing body of literature indicates that ASD severe in comorbid patients when compared to patients with only ADHD. An and ADHD traits may arise via similar cognitive processes. increased understanding of the effects of pharmacological treatment on cognitive functioning may provide more insight into the exact working mechanisms of the Treatment of ASD and ADHD pharmacological therapy in ADHD-only and comorbid patients. Thus far, little is According to the Dutch multidisciplinary guidelines for the treatment of ASD and known about the cognitive working mechanisms of the selective norepinephrine ADHD, usual care for children over the age of six consists of the prescription re-uptake inhibitor atomoxetine. The clinical trial described in chapter 7 offered of medication and/or evidence-based psychosocial interventions (e.g. parent the opportunity to examine whether a pharmacologic intervention in the training, social skills training, behavioral therapies), preferably both medication noradrenergic system hypothesized to improve symptoms of ADHD, would also and psychosocial interventions (Trimbosinstituut, 2005; Nederlands Vereniging improve inhibitory control and affect ADHD symptoms in comorbid patients. voor Psychiatrie, 2009). Pharmacological treatments of ADHD have larger effects on behavioral problems compared to psychosocial interventions, but with a Dimensional versus Categorical Models risk of side effects such as agitation, insomnia, loss of appetite, gastrointestinal Given the continuous distribution of the ASD and ADHD traits, the DSM-defined problems, and headaches (Research Units On Pediatric Psychopharmacology cut-offs between affected and unaffected may be arbitrary, and ignorant of potential (RUPP) Autism Network, 2009; 2012; for review on side effects see Hazell, difficulties in children who score just below clinical cut-offs (e.g. Constantino, 2007). Current pharmacological treatments for ASD primarily target comorbid 2011; Larsson, Anckarsater, Rastam, Chang& Lichtenstein, 2012; Levy, Hay, symptoms (e.g., irritability, aggression, hyperactivity, anxiety) rather than core McStephen, Wood & Waldman, 1997; Lundstrom et al., 2012; Plomin, Haworth social and communication impairments, while pharmacological treatment for & Davis, 2009; Robinson, Munir et al., 2011). That is, samples from the general ADHD is effective for reducing impairment associated with core ADHD symptoms populations are usually described with a lack of precision, and lumped together (i.e., inattention, hyperactivity, impulsivity) (for review, see Davis & Kollins, 2012). as an unaffected group. This overlooks the evidence that not only the ASD and This has resulted in multiple clinical trials on the effectiveness of pharmacological ADHD affected populations, but also the general population is characterized by treatment for ADHD symptoms in ASD-patients (Arnold et al., 2006; Benvenuto, behavioral, cognitive and genetic variance (e.g. Barnett, Heron, Goldman, Jones Battan, Porfirio, & Curatolo, 2012; Cortese, Castelnau, Morcillo, Roux, & Bonnet- & Xu, 2009; Constantino, 2011; Fair, Bathula, Nikolas & Nigg, 2012; Plomin et Brilhault, 2012; Doyle & McDougle, 2012; for review see Ghanizadeh, 2012; al., 2009; Robinson, Koenen et al., 2011). Barnett and colleagues (2009) for Handen, Taylor, & Tumuluru, 2011; Hanwella, Senanayake & de Silva, 2011; instance, described that a variant in the catechol-O-methyltransferase (COMT) Harfterkamp et al., 2012; Murray, 2010; Posey et al., 2007; Research Units On gene that contributes to the risk of psychiatric disorders also affects normal Pediatric Psychopharmacology (RUPP), 2005; Santosh, Baird, Pityaratstian, cognitive variation. In addition, Fair and colleagues (2012) showed that a large 24 25 Chapter 1 part of the cognitive heterogeneity within ADHD seems actually not exclusively contrast to this positive hypothesis, from an evolutionary perspective averageness related to ADHD per se, but represented cognitive heterogeneity also present may be an adaptive trade-off against the mixture of costs and benefits of more in the ‘unaffected’ (or non-symptomatic) side of the ADHD continuum. These extreme ends of the continuum (Nettle, 2006). That is, being at the lowest risk for findings point out that behavioral effects of cognitive and genetic variants are best ASD or ADHD may also come with specific disadvantages. For example, being regarded as process specific rather than disease specific. highly reflective instead of impulsive may lead to inertia, and very low levels of Empirical support is strong for an alternative model which replaces restrictive and repetitive behaviors may lead to difficulties keeping a daily routine. DSM-defined categories with multiple dimensions based on the overall number If true, the ASD and ADHD traits may be bipolar, and the most favorable trait may of symptoms present (for review see Willcutt et al., 2012). The study by Marcus be a trade-off between advantages and disadvantages of more extreme traits. To & Barry (2011) for example, compared a categorical typology with dimensional examine whether the positive or evolutionary perspective describes the ASD and measures of ADHD, and found correlations among the dimensional scores and ADHD roots best, chapter 2 examined whether the ‘lowest risk side’ of the ASD associated features consistently higher than correlations among the categorical and ADHD trait distributions presents with fewer comorbid problems and superior scores and associated features. These findings indicate that ADHD symptoms all cognitive functioning. had a dimensional structure. Likewise, quantitative symptom measures are well able to pick up on biological risk factors involved in ADHD (e.g. Bralten et al., 2013; Identifying more Homogeneous Subgroups Nigg, Goldsmith, & Sachek, 2004; Nikolas & Burt, 2010; Sonuga-Barke, 2005). Traditionally, studies used DSM-defined and therefore heterogeneous groups of Therefore, dimensional measures were analyzed in the chapters 3 to 6 of this patients. Attempts to detect shared and specific underpinnings for ASD and ADHD thesis. Furthermore, chapters 2 and 5 focused solely on the general population, may have been hindered by this heterogeneity in symptom presentation and using dimensional measures that were sensitive assessments for variance across underlying mechanisms (Bernfeld, 2012; Cross-Disorder Group of the Psychiatric the continuous ASD and ADHD traits, including the non-symptomatic ends. Genomics, 2013; for review see Hyman, 2007; Hyman, 2010; Miller, 2010; for Apart from acknowledging the continuously distributed nature of both review see Uher & Rutter, 2012). That is, specific cognitive impairments, brain ASD and ADHD within the population, thinking quantitatively may also lead to variations or genetic deficits may underlie the disorder only in a subgroup of the thinking positively, as suggested by Plomin and colleagues (2009). Instead of patients studied, and clinical diagnoses of ASD and ADHD may actually include focusing solely on the vulnerabilities of people at risk for ASD or ADHD, a new multiple, etiologically distinct subtypes with overlapping symptom presentation direction for research is to consider the potential resilience of individuals with low (e.g. Brieber et al., 2007; Buhler, Bachmann, Goyert, Heinzel-Gutenbrunner & risk rates. Individuals at the low-risk side of the ASD and ADHD continua may, Kamp-Becker, 2011; Fair et al., 2012; Gadow et al., 2009; Maher, 2008; Reiersen for example, present with excellent social skills and below average hyperactivity & Todorov, 2013; Verté, Geurts, Roeyers, Oosterlaan & Sergeant, 2006). An and impulsivity, opposite to individuals at the high-risk side of the continuum. important strategy to overcome this heterogeneity is to empirically segment this Increased understanding of potential invulnerabilities and motivational factors group of individuals with ASD symptoms, ADHD symptoms or a combination of may provide insight into mechanisms that promote favorable outcomes, which ASD and ADHD symptoms into subgroups with possibly a more homogeneous may be different from mechanisms that help to avoid unfavorable outcomes. In set of underlying mechanisms. A dimensional approach that relaxes not only the 26 27 Chapter 1 DSM-defined assumption that ASD and ADHD are two separate conditions, but Figure 1.4 Dissecting into homogeneous groups also that ASD and ADHD each consist of distinct subtypes, is warranted. Each circle represents one child, each color represents one homogeneous subgroup. Circles of the same colour form a subgroup (thus children forming classes or profiles). Based on, for example, symptom data or cognitive data, mathematical detection algorithms may detect several homogeneous subgroups with very similar profiles of behavioral and/or cognitive data. Using dimensions as the starting point, there are several methods of empirically dissecting heterogeneity and defining more homogeneous disease profiles, such as latent class analyses (LCA) or Community Detection (CD) analyses (McCutcheon, 1987; Newman, 2006). LCA areempirical bottom-up approaches that through mathematical detection algorithms can identify groups of participants (also referred to as classes or profiles) who have very similar scores on for example behavioral symptom measures (e.g. questionnaires, interviews), see also Figure 1.4. LCA result in measures of overall fit, such as the Bayesian Information Criterion (BIC) values and entropy. These measures are used as indicators of whether the data support the latent class model in question, and assess the relative efficiency with which different models fit the data (Nylund, Asparouhov & Muthén, 2007). Multiple studies disclosed clinically relevant, more homogeneous latent classes in ASD and ADHD with the use of quantitative symptom measures (e.g. Acosta et al., 2008; Constantino, 2011; Mulligan et al., 2009; Reiersen, Constantino, Volk & Todd, 2007; St. Pourcain et al., 2011; Volk, Todorov, Hay & Todd, 2009). Overall, these studies indicated that empirically defined ASD-ADHD classes derived from quantitative symptom measures show partially distinct patterns of comorbid pathology and distinct developmental trajectories. Such behaviorally homogeneous disease subtypes may be differentially linked to etiological factors, and may help reveal shared and unique mechanisms for ASD and ADHD. Therefore, LCA were used to examine etiologically different subtypes in the chapters 3 to 6 of this thesis. One step further is to base bottom-up classification on more objectively measured cognitive performances rather than on symptoms scores. As such, cognitive homogeneous subgroups related to dimensions of ASD and ADHD may bring us closer to etiological homogeneity. Thus far, research on homogeneous cognitive profiling is limited. A study of Fair and colleagues (2012) on cognitive heterogeneity in typically developing children and children with ADHD indicated that ADHD-behavior may be rooted in multiple cognitively distinct profiles. These cognitive profiles did not differ in ADHD symptom presentation, and were highly similar in the population-based and ADHD-affected samples. This may indicate that the cognitive profiles revealed through bottom-up approaches are generic, not only relevant for normal development and ADHD, but possibly also for other neurodevelopmental disorders such as ASD. Cognitive profiles may provide both research and clinical practice with cognitive substrates in children assessed for ASD and ADHD. Thus, bottom-up cognitive profiling deserves more investigation and may ultimately inform the development of more tailored diagnostic and treatment procedures. Therefore, this approach was applied in chapter 6 of this thesis. 28 29 Chapter 1 Aims and Outline of this Thesis The overall aim of this thesis is to examine shared and unique behavioral key findings from all chapters, points out limitations, suggests recommendations for future research and closes with some clinical implications. and cognitive profiles in ASD and ADHD. The approach has the following key characteristics: I) Assessing ASD and ADHD symptoms in parallel. II) Examining the relationship between behavior and cognition. III) Applying a dimensional approach, focusing on both the lower and higher end of the ASD and ADHD trait continua by integrating data from populationbased and clinic-based samples. IV) Identifying subgroups that are homogeneous at the behavioral or cognitive level, using latent class analyses. Information with regard to behavioral and cognitive functioning is collected from both general and clinic-based populations. For details on the study samples, see Box 1.2. The cognitive measures under study cover affirmed cognitive strengths and difficulties for ASD and ADHD, such as inhibition, visual and verbal attention, visual and verbal working memory, the recognition of emotions, and motor timing (for review see Rommelse et al., 2011). Chapter 2 focuses on the continuum of the ASD and ADHD traits in the general population. This study examines whether the lower end of both trait distributions represents superior functioning, or that the lowest risk for ASD and ADHD may also come with specific disadvantages. In chapters 3 to 5, reduced heterogeneity on the behavioral level aims to unravel unique and shared cognitive profiles. Next, in chapter 6 a reversed approach is used, in which a reduced heterogeneity on the cognitive level aims to detect unique and shared ASD-ADHD profiles. In chapter 7, the relationship between behavior and cognition is further explored in a double blind, placebo controlled study. This study describes whether ADHD symptom improvement is mediated by improvements in inhibitory control. Finally, chapter 8 provides a general overview, which summarizes and discusses 30 31 Box 1.2 Study Samples Schoolkids Project Interrelating DNA and Endophenotype Research (SPIDER) Eligible children were 378 children from a random population cohort, largely recruited from primary schools across the Netherlands (Breda, Enschede, Groesbeek, Hoofddorp, Malden and Nijmegen). All children were between 6 and 13 years of age (M (SD) = 8.9 (1.7)), with boys and girls equally represented (% male = 49.5). All children were of Caucasian descent and had an estimated total IQ of at least 70 on the Wechsler Intelligence Scale (WISC-III) (Wechsler, 2002). ASD, ADHD and comorbid symptoms were examined with multiple questionnaires filled in by parents. Exclusion criteria were epilepsy, known genetic or chromosomal disorders, brain damage, and problems with vision or hearing. ASD and ADHD diagnoses were not an exclusion criterion; children were asked not to use medication prior to the neuropsychological assessment. Biological Origins of Autism (BOA) study Eligible childrenwere 274 ASD (and ADHD) affected children and their siblings, recruited via Karakter Child and Adolescent Psychiatry Centers, and the Dutch organization for autism (Nederlandse Vereniging voor Autisme; NVA) .All children were between 5 and 17 years of age (M (SD) = 11.2(3.1)), with boys overrepresented(% male = 65.3). ASD, ADHD and comorbid symptoms were examined with multiple questionnaires filled in by parents; these ratings were with regard to child’s functioning off medication. All children were screened with the Social Communication Questionnaire (SCQ) (Rutter et al., 2003) completed by parents and teachers. To avoid false negatives, families were included if at least one child presented a score above 10 on the parent version or above 15 on the teacher version of the SCQ. For all children scoring above the cut-off, a formal diagnosis of ASD was made by a certified clinician using the Autism Diagnostic Interview-revised (ADI-R) (Le Couteur, Lord, & Rutter, 2003). The Conners’ Rating Scales-Revised (CRS-R) (Conners, Sitarenios, Parker & Epstein, 1998a; 1998b) completed by parents and teachers was used to screen for ADHD. A T-score above 63 on one of the ADHD-subscales of the CRS-R was considered clinical. For all children scoring above this cut-off, or previously having a diagnosis of ADHD, the parental account for childhood symptoms (PACS) was administered by a certified clinician to obtain a diagnosis of ADHD (Taylor, Sandberg, Thorley, & Giles, 1991). Further inclusion- and exclusion criteria were identical to those listed for the SPIDER. Total IQ was estimated via the Wechsler Preschool and Primary Scale of Intelligence (WPPSI), Wechsler Intelligence Scale (WISC-III) or Wechsler Adult Intelligence Scale (WAIS-III) (Wechsler 1989; 2000; 2002). Children were asked not to use medication prior to the neuropsychological assessment. Research on Atomoxetine in Dutch ASD/ADHD Children (RADAR) Eligible childrenwere 97 ASD and comorbid ADHD affected children, referred to one of nine participating child and adolescent psychiatry centers across the Netherlands (Amsterdam, Groningen, Hoorn, Leiden, Maastricht, Nijmegen, Oosterhout, The Hague and Utrecht). All children were between 6 and 17 years of age (M (SD) = 10.4 (2.9)), with boys overrepresented (% male = 85.6), and had an estimated total IQ of at least 60 on the Wechsler Intelligence Scale (WISC-III) (Wechsler, 2002). Exclusion criteria were a weight of less than 20 kg, presence of psychosis, bipolar disorder, substance abuse, a serious medical illness, history of seizures, ongoing use of psychoactive medications other than the study drug, and intended start of a structured psychotherapy or in-patient treatment. Females who were post-menarche and sexually active had to take a pregnancy test to exclude pregnancy. Clinical Trial Registry B4Z-UT-S017, NCT00380692. 32 33 Are high and low extremes of ASD and ADHD trait continua pathological? A population-based study using the AQ and SWAN rating scales Jolanda M. J. van der Meer*, Corina U. Greven*, Catharina A. Hartman, Martijn G. A. Lappenschaar, Jan K. Buitelaar, Nanda N. J. Rommelse *joint first author Under review 34 Abstract Chapter 2 Autism Spectrum Disorder (ASD) and Attention-Deficit/Hyperactivity Disorder (ADHD) are thought to exist on a continuum, where diagnosis simply reflects the symptomatic end of a normal distribution of quantitative traits in the general population. One implicit assumption is that the non-symptomatic ends of these Are high and low extremes of ASD distributions represents superior functioning; however, this remains to be tested, as it is possible that symptomatic and non-symptomatic ends of clinical and ADHD trait continua pathological? continua are both pathological. Data came from 378 children (6-13 years) from A population-based study using the AQ a population-based sample. Parents rated their child along the ASD and ADHD and SWAN rating scales trait continua using the Autism Spectrum Quotient (AQ) and the Strength and Weaknesses of ADHD Symptoms and Normal behavior (SWAN) questionnaires, which show a normal distribution of scores in population-based samples. Scores on the AQ and SWAN were related to measures of internalizing and externalizing behavioral problems and cognitive functioning using polynomial regression analyses. Associations between the ASD and ADHD traits on the one hand, and behavioral problems and cognitive functioning on the other hand, were largely linear. The non-symptomatic ends of the ASD and ADHD trait continua were not pathological; instead, they were on average related to fewer behavioral problems and better cognitive functioning than the symptomatic ends. Finding linear relations suggests that the non-symptomatic ends of the ASD and ADHD trait continua differ largely quantitatively rather than qualitatively. Studying the correlates of the non-symptomatic ends of these continua may deepen our understanding of the mechanisms underlying superior behavioral and cognitive functioning. Jolanda M. J. van der Meer*, Corina U. Greven*, Catharina A. Hartman, Martijn G.A. Lappenschaar, Jan K. Buitelaar & Nanda N.J. Rommelse *joint first author Under review 37 Chapter 2 Psychiatric disorders such as Autism Spectrum Disorder (ASD) and Attention- These measures, however, attenuate variability at the low, non-symptomatic Deficit/Hyperactivity Disorder (ADHD) are considered to represent the extreme side of the distribution where individuals are grouped together into a single no- problematic manifestations of a continuous distribution of quantitative traits in symptoms group as reflected in the highly skewed distribution of scores of such the general population. Evidence for this comes from a large body of research measures (Delucchi & Bostrom, 2004). Two more recently developed measures using a complementary array of research design, methodological and statistical that cater for this methodological limitation are the Autism Spectrum Quotient approaches (Constantino, 2011; Widiger & Samuel, 2005; Willcutt, 2005). An (AQ) and the Strength and Weaknesses of ADHD Symptoms and Normal behavior underlying assumption to viewing ASD and ADHD as symptomatic extremes of (SWAN) questionnaires (Auyeung, Baron-Cohen, Wheelwright & Allison, 2008; quantitative trait continua is that the non-symptomatic ends of these continua Hay, Bennett, Levy, Sergeant & Swanson, 2007). The AQ and SWAN both result represent superior functioning, or possibly even resilience rather than only in a continuous distribution of scores in the general population (Hoekstra, Bartels, indicating low risk (Plomin et al. 2009; 2012). However, this assumption may not Cath, & Boomsma, 2008; Polderman et al., 2007) and can therefore be regarded necessarily be correct. to measure individual differences in ASD and ADHD traits on a continuum from low Being at the lowest risk for ASD or ADHD may also come with specific to high. It has been shown that children who show no variation on conventional disadvantages (Plomin et al., 2009). For example, being highly focused in terms skewed scales of ADHD behaviors, show variation across the non-symptomatic of attention may lead to less flexible shifting of attention and cognitive rigidity, side of the SWAN scale (Arnett et al., 2011; Polderman et al., 2007). being highly reflective instead of impulsive and able to control activity levels may lead to lack of spontaneity and inertia. Likewise, the opposite end of the years) found evidence suggesting that the non-symptomatic end of the ADHD social communication and interaction deficits seen in ASD may reflect other distribution is indeed linked to superior functioning rather than representing types of abnormal social behavior (e.g., very high empathy levels could also be psychopathology (Crosbie et al., 2013). The study showed that individuals impairing), and very low levels of restrictive and repetitive behaviors could come with the highest possible ADHD traits performed worse on the stop signal task with difficulties keeping a daily routine and making plans. Thus, the ASD and (assessing response inhibition, response latency and response variability) than ADHD traits may be bipolar, where both extremes represent maladaptation and those with the lowest possible ADHD traits (Crosbie et al., 2013). Moreover, parent psychopathology. In that case, the most favorable trait may be a trade-off between reports of ADHD, anxiety, depression, learning disability and other disorders were advantages and disadvantages of more extreme traits, and superior functioning lowest for participants with the lowest ADHD trait scores, althoughthe study did will be associated with people at the trade-off level (Nettle, 2006; Plomin et al., not present relevant statistics in support of this conclusion. It remains to be tested 2009). whether results replicate for cognitive domains associated with ADHD other than Viewing ASD and ADHD as symptomatic extremes of quantitative traits A recent study using the SWAN in a community sample (ages 6-18 those assessed by the stop task, and for behavioral problems other than ADHD. calls for a perspective and methodology different from the traditional categorical focus described in current psychiatric classification schemes. Typically, interview whether the non-symptomatic ends of the ASD and ADHD trait continua are and questionnaire measures that assess complex psychiatric disorders allow fine- associated with fewer internalizing and externalizing behavioral problems scaled assessment of variability at the high, symptomatic end of the distribution. and better cognitive functioning. This question is addressed using polynomial 38 The present study uses the AQ and SWAN questionnaires to examine 39 Chapter 2 regression analyses in a population-sample of 378 children. The study focuses The SWAN consists of 18 DSM-IV-based items scored on a 7-point Likert scale on the ASD and ADHD continua as they refer to two commonly associated from (1) ‘far below’, (2) ‘below’, (3) ‘slightly below’, (4) ‘average’, (5) ‘slightly complex disorders and traits that share comorbidity patterns with other behavioral above’, (6) ‘above’ to (7) ‘far above’ (possible range of scores: 18-126; actual problems and which are assumed to share cognitive underpinnings (see also van range in the present sample: 25-118). Each item on the SWAN is phrased to der Meer et al. (in press) in chapter 5; for review see Rommelse et al., 2011). represent behavior across the ADHD continuum rather than to represent a symptom, thereby defining a middle range of behavior. Ratings are then made in METHODS relation to deviation from this middle range. Scores on the SWAN were mirrored Participants to match the directionality of the AQ: higher scores were indicative of more ASD Participants were part of a population based study of children sampled from primary schools across the Netherlands (the Schoolkids Project Interrelating DNA and Endophenotype Research – SPIDER) Data collection took place between January 2009 and July 2011. The study sample consisted of 378 children of Caucasian descent between the ages of 6 and 13 years (M(SD) = 8.9(1.7), % males = 49.5). Exclusion criteria were epilepsy, known genetic or chromosomal disorders, brain damage, problems with vision or hearing, and an estimated total IQ of less than 70 on the Wechsler Intelligence Scale (WISC-III) (Wechsler, 2002). and ADHD traits. Both measures have adequate reliability and validity (Arnett et al., 2011; Hoekstra et al., 2008), and demonstrated sensitivity and specificity to identify clinically elevated scoring children (Arnett et al., 2011; Auyeung et al., 2008). In the current study the two measures correlated .26 (p < .001), and showed a gamma distribution with a large shape parameter (see Supplement 2.1), which closely resembles a normal distribution. Internalizing and externalizing behavioral problems Informed written consent was obtained from parents. SPIDER is approved by the Parents also rated their child’s behavior on pencil-and-paper copies of the Committee on Research involving Human Subjects (CMO). Conners’ Parent Rating Scale-Revised: Long version (Conners et al., 1998a). The Conners’ scale is a validated questionnaire for assessing behavioral problems Measures including oppositional behavior, emotional lability, anxiety, perfectionism and ASD and ADHD traits psychosomatic complaints. Parents were invited to rate their child’s ASD and ADHD traits using pencil-andpaper copies of the AQ (Auyeung et al., 2008; Hoekstra et al., 2008) and SWAN (Hay et al., 2007) questionnaires. The AQ consists of 50 items scored on a 4-point Likert scale from (0) ‘definitely agree’, (1) ‘slightly agree’, (2) ‘slightly disagree’ to (3) ‘definitely disagree’ (possible range of scores: 0-150; actual range in the present sample: 8-120). Around half of the items on the AQ are phrased as to elicit a positive endorsement from individuals high in ASD traits (e.g., ‘I am fascinated by dates’), the other half a negative endorsement (e.g., ‘I find social situations Cognitive functioning The cognitive tasks included in this study are briefly described in Table 2.1. The cognitive domains assessed by these tasks have previously shown to be impaired in both ASD and ADHD (Oerlemans et al., 2013; Rommelse et al., 2011; van der Meer et al., 2012, as described in chapter 3). No ceiling effects occurred. Scores on some of the cognitive tasks were mirrored, so that higher scores on all cognitive variables implied better performance. easy’), which results in a continuous distribution of scores in population samples. 40 41 Chapter 2 Table 2.1 Description of the cognitive tasks Data analyses Task Measurement potential Dependent variable(s) Polynomial regression analyses were conducted which included the linear and Baseline Speed a,b Speed and variability of motor output as response to external cue MRT (ms) and variability (SD of reaction time in ms) quadratic terms of the ASD or ADHD traits as predictors, and the internalizing Facial Emotion Recognitiona,b Capacity to identify the facial emotional expression of happiness, sadness, anger and anxiety MRT (ms) and accuracy on four emotions Linear relations between predictors and outcome measures would indicate Visuo-Spatial Sequencinga,b Visuo-spatial attention Number of correctly reproduced sequences in identical (forward) order relations, evidenced by significant quadratic terms, could suggest that the non- Visuo-spatial working memory Number of correctly reproduced sequences in reversed (backward) order Centering of predictors was used (Bradley & Srivastava, 1979). All outcome Verbal attention Number of correctly reproduced digits in identical (forward) order 1975) and standardized into z-scores (SPSS version 18). As an exploratory step, Verbal working memory Number of correctly reproduced digits in reversed (backward) order effects of sex and age were also examined. Block Patterns (WISC-III)a,c Visual pattern recognition Number of correctly and timely completed geometric designs Response Organization Objectsa,b Motor inhibition Difference in percentage of errors or MRT (ms) between compatible and incompatible trials ADHD-term, to examine prediction beyond age and sex effects (step 1 linear). The Difference in percentage of errors or MRT (ms) between compatible trials and mixed compatible-incompatible trials. (ASD*age, ASD*sex; or ADHD*age, ADHD*sex; step 2 linear). The alternative Digit Span (WISC-III)a,c Cognitive flexibility and externalizing behavioral problems and cognitive tasks as outcome measures. that the non-symptomatic ends of the ASD and ADHD trait continua indeed represent better functioning than the symptomatic ends. In contrast, curvilinear Note. MRT=mean reaction time, a van der Meer et al.(2012); b de Sonneville (1999) ;c WISC-III = Wechsler Intelligence Scale (Wechsler, 2002). symptomatic ends may be pathological, depending on the shape of the curve. measures were normalized using a Van der Waerden transformation (Lehman, The basic regression model included only age and sex as predictors (step 0). The first extended model covered this basic model plus the linear ASD or second extended model also added interactions with age and sex as predictors second extended model also added the quadratic term (ASD2 or ADHD2; step 2a quadratic) and interactions with age and sex as predictors (ASD2*age, ASD2*sex; or ADHD2*age, ADHD2*sex; step 2b quadratic). The False Discovery Rate (FDR) controlling procedure was applied to correct for possible multiple testing effects (p-value set at 0.05)(Benjamini, 1995).For illustrative purposes, scatterplots were created for any analyses that revealed curvilinear effects, to examine the shape of the quadratic curve. As a post-hoc step which further confirmed impressions from regression analyses and scatterplots, children with the highest and lowest scores on the ADHD and ASD distributions were compared on mean internalizing and externalizing behavioral problems and cognitive functioning (for a description of these analyses see Supplements 2.2 and 2.3). 42 43 Chapter 2 RESULTS of the interactions between the linear or quadratic ADHD terms and age were The ASD trait significant. The linear ASD term was significantly associated with increased scores on all internalizing and externalizing behavioral problems (step 1 linear, Table 2.2), but with none of the cognitive variables (step 1 linear, Table 2.3). The quadratic ASD term was not associated with any of the behavioral problems (step 2a-b quadratic, Table 2.2). However, the interaction between quadratic ASD term and sex was a significant predictor of visuo-spatial working memory (step 2b quadratic, 2.3). Post-hoc testing revealed that the quadratic ASD term was a significant predictor of visuo-spatial working memory in girls (β=-0.33, p<.001), but not boys (β=0.09, non-significant). No other associations between the quadratic ASD term and the cognitive variables were significant. In addition, none of the interactions between the linear or quadratic ASD terms and age were significant. The ADHD trait The linear ADHD term was significantly associated with increased scores on oppositional behavior, emotional lability, anxiety and psychosomatic complaints, but not with perfectionism (step 1 linear, Table 2.2).In addition, theinteraction between the linear ADHD term and sex was a significant predictor of oppositional behavior and perfectionism (step 2 linear, Table 2.2). Post-hoc testing indicated that oppositional behavior and psychosomatic complaints were more strongly related to the linear ADHD term in boys (β=0.56, p<0.001; and β=0.44, p<0.05, respectively) than girls (β=0.20, p<0.01; and β=0.16, p<0.05, respectively). The linear ADHD term was also significantly associated with worse performance on visuo-spatial and verbal working memory and visual pattern recognition (step 1 linear, Table 2.3), but not with any other cognitive measures. The quadratic ADHD term showed significant associations with oppositional behavior, emotional lability and perfectionism (step 2a quadratic, Table 2.2). No significant associations were found between the quadratic ADHD term and any of the cognitive variables (step 2a-b quadratic, Table 2.3). None 44 45 46 Predictor - SWAN2*sex ∆R2=0.01 ∆R2=0.03 ∆R2=0.03 - - 0.19 (0.00) [0.00,0.00] - - 0.32 (0.02) [0.01,0.03] 2 ∆R2=0.01 ∆R2=0.03 ∆R2=0.01 ∆R2=0.10 ∆R2=0.01 ∆R2=0.00 ∆R2=0.00 ∆R2=0.15 R2=0.00 R /∆R 2 - - - - - 0.21 (0.01) [0.01,0.02] - - - - - 0.48 (0.02) [0.02,0.03] - - β (B) [95% C for B] 2 R /∆R 2 ∆R2=0.00 ∆R2=0.01 ∆R2=0.01 ∆R2=0.04 ∆R2=0.01 ∆R2=0.01 ∆R2=0.01 ∆R2=0.21 R2=0.02 Anxiety - - 0.18 (0.00) [0.00,0.00] - - - - - - - - 0.50 (0.03) [0.02,0.03] -0.20 (-0.37) [-0.56,-0.19] - 2 ∆R2=0.00 ∆R2=0.03 ∆R2=0.01 ∆R2=0.00 ∆R2=0.02 ∆R2=0.01 ∆R2=0.00 ∆R2=0.23 R2=0.04 R /∆R 2 Perfectionism β (B) [95% C for B] Internalizing and externalizing problems - - - -0.20 (-0.02) [-0.04,-0.01] - 0.32 (0.02) [0.01,0.03] - - - - - 0.23 (0.01) [0.01,0.02] - - β (B) [95% C for B] ∆R2=0.01 ∆R2=0.01 ∆R2=0.02 ∆R2=0.10 ∆R2=0.00 ∆R2=0.00 ∆R2=0.00 ∆R2=0.05 R2=0.00 R2/∆R2 Psychosomatic Complaints Note. AQ=Autism Spectrum Quotient scale .SWAN=Strengths and Weaknesses of ADHD symptoms and Normal behavior scale. β=standardized regression coefficient, B=unstandardized regression coefficient (with 95% confidence interval). Sex defined as boys=0 and girls=1. R 2=proportion of variance explained by variables in step 0. ∆R 2=proportion of variance explained by the additionally entered predictors. Step 2 (linear) and steps 2a-b (quadratic) are alternative steps, not consecutive steps. Regression coefficients shown only for new predictors entering at each step. A higher score on the AQ, SWAN, and internalizing and externalizing problems indicated more behavioral problems. Hence, a positive β(B) value indicates that an increase in ASD/ADHD traits relates to an increase in internalizing and externalizing problems. Values printed in bold are significant at p<0.05. β(B) values are only reported if significant after FDR correction. - 0.19 (0.00) [0.00,0.00] SWAN2*age Step 2b (quadratic) for ADHD SWAN 2 Step 2a (quadratic) for ADHD -0.23 (-0.02) [-0.04,-0.01] SWAN*age 0.39 (0.03) [0.02,0.03] ∆R2=0.15 - - - - 0.40 (0.02) [0.01,0.02] - - β (B) [95% C for B] - ∆R2=0.03 ∆R2=0.00 ∆R2=0.01 ∆R2=0.15 R2=0.01 2 R /∆R 2 Emotional lability - SWAN*sex Step 2 (linear) for ADHD SWAN Step 1 (linear) for ADHD AQ2*sex AQ2*age Step 2b (quadratic) for ASD AQ2 - - AQ*sex Step 2a (quadratic) for ASD - AQ*age Step 2 (linear) for ASD AQ 0.40 (0.02) [0.02,0.03] - sex Step 1 (linear) for ASD - age Step 0 (same for ASD and ADHD) β (B) [95% C for B] Oppositional behavior Table 2.2 Polynomial regressions of the ASD (AQ) and ADHD (SWAN) trait measures on internalizing and externalizing behavioral problems Chapter 2 47 Chapter 2 Table 2.3 Polynomial regressions of the ASD (AQ) and ADHD (SWAN) trait measures on cognitive tasks Cognitive tasks Working memory Visuo-spatial β (B) [95% C for B] Block patterns Verbal β (B) [95% C for B] Visual patterns R2/∆R2 β (B) [95% C for B] R2/∆R2 Predictor Step 0 (same for ADHD and ASD) R =0.38 R =0.19 2 R =0.03 2 2 age 0.61 (0.34) [0.30,0.39] 0.43 (0.24) [0.19,0.29] -0.16 (-0.09) [-0.14,-0.03] sex - - - ∆R2=0.00 Step 1 (linear) for ASD AQ - Step 2 (linear) for ASD ∆R2=0.00 - - - - - - - ∆R2=0.00 - Step 2b(quadratic) for ASD ∆R2=0.00 - ∆R =0.00 - - - AQ2*sex -0.27 (-0.00) [-0.00,-0.00] - - SWAN ∆R2=0.02 -0.15 (-0.01) [-0.02,-0.00] ∆R2=0.02 -0.13 (-0.01) [-0.01,-0.00] ∆R2=0.00 Step 2 (linear) for ADHD ∆R2=0.05 - - - SWAN*sex - - - Step 2a(quadratic) for ADHD SWAN2 ∆R =0.00 - ∆R =0.00 - ∆R2=0.00 Step 2b(quadratic) for ADHD - - - - SWAN *sex - - - 2 Note. For a description of the table contents, see also the note to Table 2.2. A higher score on the cognitive tasks indicated better performance. Hence, a negative β(B) value indicates that an increase in ADHD/ASD traits relates to a decrease in cognitive functioning. Results for the other cognitive tasks included in this study (the Baseline Speed, Facial Emotion Recognition and Response Organization Objects tasks, as well as visuo-spatial and verbal attention on the Visuo-Spatial Sequencing and Digit Span tasks) are not tabulated, as neither the linear nor quadratic ASD or ADHD terms were significant predictors of these tasks (i.e., steps 1, 2, 2a and 2b yielded non-significant results). 48 previous evidence showing that impairments in spatial working memory are linked to ADHD, but not ASD symptoms (Sinzig, Morsch, Bruning, Schmidt, & Lehmkuhl, ∆R2=0.00 SWAN2*age complaints). In addition, the ADHD but not the ASD trait was significantly and verbal and visuo-spatial working memory. The latter finding is consistent with 2 ∆R2=0.00 significantly associated with internalizing and externalizing behavioral problems associated with reduced cognitive performance on visual pattern recognition ∆R =0.00 2 symptomatic ends (Plomin et al., 2009; 2012), or also represent maladaptive (oppositional behavior, emotional lability, anxiety, perfectionism, psychosomatic ∆R2=0.01 SWAN*age This study examined if the non-symptomatic ends of the ASD and ADHD trait outcomes (Nettle, 2006; Plomin et al., 2009). The ASD and ADHD traits were -0.23 (-0.02) [-0.02,-0.01] ∆R2=0.00 2 behavioral problems and better cognitive functioning than the symptomatic ends continua are associated with better behavioral and cognitive outcomes than the 2 AQ2*age Step 1 (linear) for ADHD associated with the symptomatic end. Hence, for any outcome measures showing DISCUSSION ∆R2=0.00 ∆R =0.01 ∆R =0.02 ADHD scores from which children improved at a slower rate on negative outcomes (see Supplements 2.4 and 2.5). 2 2 symptomatic end, a point was reached at more intermediate levels of ASD or ∆R2=0.01 AQ*sex AQ2 than a u-shaped distribution: moving down from the symptomatic to the non- non-symptomatic end of the ASD and ADHD distributions were linked to fewer - AQ*age Step 2a(quadratic) for ASD The nature of curvilinear relations was best described in terms of a j- rather significant associations with the linear or quadratic ASD or ADHD terms, the ∆R2=0.00 ∆R2=0.01 ∆R2=0.01 Illustration of curvilinear relationships 2008). Any significant associations between the ASD or ADHD trait on the one hand, and behavioral problems or cognitive functioning on the other hand were linear, with the exception of oppositional behavior and emotional lability, which showed curvilinear (quadratic) relations with the ADHD trait in addition to linear associations, and perfectionism which only showed a curvilinear association with the ADHD trait. There was some evidence for sex effects. Oppositional behavior and psychosomatic complaints were more strongly related to the ADHD trait in 49 Chapter 2 boys, and a curvilinear relation between the ASD trait and visuo-spatial working are likely to be larger (Preacher, Rucker, MacCallum & Nicewander, 2005). memory was only significant in girls. Curvilinear relations could suggest that the Normally distributed measures such as the AQ and SWAN could for example be non-symptomatic ends represent maladaptation, depending on the shape of helpful in selecting control individuals with particularly low ASD or ADHD scores the quadratic curve. Given the j-shaped curve, for any behavioral and cognitive in case-control studies, to increase power if effect sizes are expected to be small. outcomes showing significant linear or curvilinear associations with the ASD However, these effects can only be relevant across the distribution of traits if or ADHD trait, the non-symptomatic ends of the trait continua were associated high and low extremes differ quantitatively rather than qualitatively, as tentatively with fewer behavioral problems and better cognitive performance than the shown here for the first time for the ASD and ADHD traits. symptomatic ends. In contrast, neither linear nor curvilinear associations were found between the included internalizing and externalizing behavior scales were traditional the ADHD trait and other included cognitive measures (i.e. baseline speed, skewed measures, which show little variation at the non-symptomatic side of facial emotion recognition, attention, motor inhibition and cognitive flexibility). the distribution. Although the non-symptomatic ends of the ASD and ADHD- Moreover, the curvilinear association between the ASD trait and visuo-spatial trait continua were associated with lower levels of behavioral problems, it is working memory in girls represented the only significant prediction of cognitive unclear whether this extends to the presence of more positive behaviors or even functioning from the ASD trait, and hence interpretive caution is warranted. resilience at the non-symptomatic ends. The current dataset contained no suitable Together these findings suggest that most of the cognitive functioning is stable measures to examine whether the non-symptomatic ends are associated with across the ASD and ADHD trait continua in the general child population. Since positive constructs such as wellbeing or quality of life and this therefore remains associations with the cognitive measures were almost exclusively limited to the an interesting question for future research. Nonetheless, the included internalizing ADHD trait, the present regression analyses provide little evidence in support of and externalizing questionnaires represent gold-standard measures, and to the the hypothesis that ASD and ADHD traits share cognitive underpinnings, as was best of our knowledge alternative continuously distributed behavior problem outlined elsewhere (van der Meer et al., 2012, see also chapter 3; Rommelse et measures do not exist. Importantly, the present study revealed a consistent pattern al., 2011). of association at the extremes. What explains this pattern merits examination in Findings from this study provide further support for the assumption future research. Second, the present study is based on a moderate-size population that ASD and ADHD are extreme manifestations of quantitative traits that cover sample of typically developing children. The study questions raised in this paper quantitative rather than qualitative differences. In line with research conducted by should be replicated in a larger normative epidemiological sample. Third, only Fair and colleagues (2012) and the study described in chapter 5 (van der Meer some of the associations between the ASD and ADHD traits and the cognitive et al., in press), heterogeneity in ASD and ADHD may be rooted in heterogeneity tasks were significant. Although using a sample of typically developing children is in the non-symptomatic end of the trait distributions. These findings may have also a strength, it may have weakened associations between the ASD and ADHD implications for selecting individuals with high and low extreme scores on these traits and the outcome measures through attenuation of range of scores at the continua in extreme group comparison designs. Comparisons between extreme clinical extreme. Nonetheless, the present study was able to show that the ADHD groups can lead to increased power and larger effect sizes as group differences trait is linked to worse visual pattern recognition and verbal and visuo-spatial 50 This study comes with some limitations and considerations. First, 51 Chapter 2 working memory, and the ASD trait to better visuo-spatial working memory in girls. Supplemental Material Fourth, it is unclear whether the AQ and SWAN measures successfully capture the extreme of the far end of the ASD and ADHD traits, and further validation work in this area is necessary to which the current study makes a first contribution. Supplement 2.1 Gamma distribution of the ASD (AQ) and ADHD (SWAN) trait measures Fifth, the study exclusively relied on parent report of child behavior, and findings should be replicated for other informants. In conclusion, the present study suggests that the non-symptomatic ends of the ASD and ADHD trait continua are not pathological, but represent opposite ends of the ASD and ADHD trait continua appear to represent largely quantitative rather than qualitative differences. Expected gamma value better cognitive and behavioral functioning than the symptomatic ends. The Expected gamma value Total score AQ Total score SWAN Note.The AQ and SWAN scales showed a gamma distribution with a large shape parameter. This closely resembles a normal distribution, however, the Q-Q plot significantly deviated from normality. 52 53 54 1.82 1.00 378 0.00 0.94 18 1.06 1.30 19 0.60 1.02 38 0.31 0.88 38 0.42 0.99 38 0.18 0.82 75 -0.20 0.92 38 -0.25 0.86 38 -0.37 0.93 38 -0.15 0.89 19 -0.33 0.68 19 -0.76 1.77 1.00 378 0.00 0.81 18 1.07 0.84 19 0.69 0.96 38 0.55 0.97 38 0.41 0.96 38 0.15 0.91 75 -0.12 0.86 38 -0.25 0.86 38 -0.28 0.84 38 -0.51 0.93 19 -0.65 0.74 19 -0.70 1.79 1.00 378 0.00 0.85 18 1.05 0.85 19 0.96 0.94 38 0.53 0.86 38 0.56 0.96 38 0.12 0.87 75 -0.22 1.00 38 -0.14 0.69 38 -0.39 0.89 38 -0.47 0.68 19 -0.74 0.62 19 -0.74 0.96 1.00 378 0.00 1.14 18 0.62 1.22 19 0.41 1.07 38 0.10 1.01 38 0.22 0.96 38 0.22 0.88 75 -0.19 1.03 38 -0.14 0.93 38 0.00 0.84 38 -0.25 1.04 19 -0.23 0.87 19 -0.34 - 1.00 368 0.00 1.10 18 0.11 1.00 18 -0.20 1.02 36 -0.01 0.95 38 0.04 0.88 37 -0.12 0.97 74 0.16 0.99 37 -0.16 1.25 37 -0.07 1.01 36 0.08 1.00 19 0.05 0.80 18 -0.14 MRT speed - 1.00 363 0.00 0.98 18 0.29 0.85 18 -0.10 1.07 35 0.00 1.22 38 -0.02 1.00 37 -0.01 0.96 74 0.09 0.85 36 -0.26 0.87 35 -0.17 1.02 35 0.07 1.18 19 0.33 0.89 18 -0.17 Variability Baseline Speed MRT - 0.99 185 0.00 1.08 8 0.51 0.79 10 -0.17 0.86 19 -0.27 1.13 23 0.03 0.76 16 - 0.99 185 0.00 0.51 8 -0.33 1.15 10 -0.79 1.18 19 0.58 0.89 23 -0.15 0.72 16 0.18 0.94 0.92 0.14 40 -0.17 0.91 15 -0.31 0.90 14 0.64 0.98 19 0.29 0.92 10 -0.05 1.06 11 -0.23 40 0.02 1.34 15 0.08 1.13 14 0.03 0.92 19 -0.29 1.13 10 0.42 0.96 11 -0.12 Accuracy Facial Emotion Recognition - 1.00 373 0.00 0.97 18 0.13 0.88 19 -0.12 0.69 38 0.12 1.15 38 0.00 0.95 38 0.14 1.00 74 -0.08 0.86 35 0.05 1.09 37 0.04 1.27 38 -0.08 0.69 19 -0.21 1.23 19 -0.03 - 1.00 374 0.00 1.17 18 0.10 0.80 19 -0.10 1.15 37 -0.16 0.92 38 0.00 1.07 38 0.10 0.92 73 0.06 1.09 37 0.20 0.90 38 0.12 0.94 38 -0.02 1.19 19 -0.23 0.87 19 -0.47 Visuospatial Verbal Attention 2.35a 1.00 176 0.00 0.69 2 -1.98 0.49 5 -0.29 0.94 12 -0.01 0.77 12 0.14 0.94 17 0.08 1.01 38 0.09 1.29 25 -0.04 1.05 23 -0.21 0.88 20 0.11 0.69 10 -0.15 0.94 12 0.37 1.00 - - 377 0.00 1.00 18 0.28 1.07 19 -0.19 0.85 38 -0.10 1.02 38 -0.04 1.04 38 0.02 1.01 74 0.20 1.12 38 -0.02 0.83 38 -0.20 1.14 38 0.08 0.77 19 -0.29 0.98 19 - 1.00 366 0.00 0.74 18 0.04 1.12 17 -0.20 0.78 35 0.09 0.96 37 0.24 1.08 38 -0.08 1.03 74 -0.03 0.94 36 -0.14 1.07 37 0.01 1.04 37 0.14 1.04 19 0.01 1.15 18 -0.28 - 1.00 367 0.00 0.93 18 -0.16 1.05 17 -0.21 1.28 35 -0.04 0.94 38 0.02 0.98 38 -0.02 0.94 74 0.09 1.05 36 0.25 0.99 37 0.10 0.81 37 -0.29 0.94 19 0.02 1.16 18 -0.06 Response Accutime racy Visual patterns -0.06 Motor inhibition Block patterns 1.00 374 0.00 1.00 18 0.23 1.01 19 -0.18 1.07 37 -0.04 0.85 38 -0.13 0.94 38 -0.11 1.00 73 0.17 0.86 37 -0.03 1.10 38 -0.02 1.22 38 -0.04 0.86 19 0.19 0.98 19 -0.16 Visuospatial Verbal Working memory Cognitive tasks - 1.00 368 0.00 1.02 18 0.04 1.17 18 -0.58 0.95 36 0.24 0.85 38 0.17 1.08 37 0.07 1.04 74 -0.02 0.92 36 0.06 1.02 37 -0.14 1.02 37 0.18 0.81 19 -0.40 0.90 18 -0.10 - 1.00 368 0.00 0.97 18 -0.15 0.99 18 0.01 1.19 36 0.09 0.98 38 0.37 1.04 37 0.00 1.03 74 0.03 0.79 36 0.16 0.88 37 -0.16 0.91 37 -0.24 1.01 19 -0.21 1.11 18 -0.20 Response Accutime racy Cognitive flexibility Note. AQ=Autism Spectrum Quotient scale. M=mean. SD=standard deviation. N=number of individuals. MRT=mean reaction time. ASD quantiles refer to the 5%, 10% or 20% quantiles on the AQ questionnaire. The 95-100% quantile includes the children with the highest most symptomatic scores on the AQ, the 0-5% quantile children with the lowest least symptomatic scores. All measures were corrected for age and sex. Internalizing and externalizing problems and cognitive tasks were normalized using a van der Waerden transformation and transformed into z-scores to achieve a mean of 0 and a standard deviation of 1 (see row labeled: Total [0-100%]). Through comparisons of quantiles at the top and bottom extremes it was possible to examine whether the non-symptomatic ends of the ASD distribution were on average linked to better behavioral and cognitive outcomes than the symptomatic ends. The 95-100% quantile includes the children with the highest most symptomatic scores on the AQ, the 0-5% quantile children with the lowest least symptomatic scores. SMD (standardized mean difference) refers to a comparison between children at the extremes comparing the 0-5% versus 95-100% quantile, expressed as a difference in means based on a standard deviation of 1. a There was a significant association between the quadratic AQ term and visuo-spatial working memory that was only significant in girls, but not boys (see main manuscript). Hence, mean performance on visuo-spatial working memory per quantile is tabulated for girls only. 1.89 1.00 SD SMD 378 0.00 N 1.06 SD M 18 (0-100%) 1.01 1.04 SD N 19 M 0.42 0.88 SD N 38 M 0.46 0.85 SD N 38 M 0.46 0.92 SD N 38 M -0.08 0.94 SD N 75 M -0.12 0.90 SD N 38 M -0.13 0.79 SD N 38 M -0.17 0.96 SD N 38 M -0.35 0.97 SD N 19 M -0.41 0.97 SD N 19 N M -0.88 M Total 95-100% 90-95% 80-90% 70-80% 60-70% 40-60% 30-40% 20-30% 10-20% 5-10% 0-5% ASD quantiles EmoOppo- tional Perfect- Psychositional lability Anxiety ionism somatic Internalizing and externalizing problems Supplement 2.2 Mean internalizing and externalizing behavior and mean cognitive task scores per ASD (AQ) trait quantile Chapter 2 55 56 1.76 1.00 378 0.00 1.01 18 1.25 1.11 19 0.90 0.96 38 0.18 0.84 38 -0.03 0.99 38 0.03 0.97 75 -0.23 0.95 38 -0.11 0.80 38 -0.03 0.88 38 -0.19 0.70 19 -0.39 0.98 19 -0.51 1.57 1.00 378 0.00 0.96 18 0.99 1.04 19 0.36 0.97 38 0.22 0.99 38 -0.07 1.10 38 -0.38 0.98 75 -0.07 0.89 38 0.05 0.96 38 0.04 0.78 38 -0.03 1.12 19 -0.13 0.70 19 -0.58 0.79 1.00 378 0.00 1.26 18 0.68 1.13 19 0.28 0.88 38 0.00 1.04 38 -0.07 0.96 38 -0.47 0.96 75 0.00 0.84 38 -0.02 0.99 38 -0.04 0.84 38 0.04 1.02 19 0.31 1.21 19 -0.11 1.60 1.00 378 0.00 1.20 18 0.97 1.23 19 0.51 0.91 38 0.22 1.03 38 0.21 0.75 38 0.04 0.97 75 0.03 0.87 38 -0.17 0.94 38 -0.25 0.83 38 -0.41 0.94 19 -0.18 0.94 19 -0.63 - 1.00 368 0.00 0.97 18 -0.01 1.11 17 0.05 0.98 37 -0.26 0.99 37 -0.07 0.84 35 - 1.00 363 0.00 1.04 18 -0.18 1.02 17 -0.05 0.91 37 -0.17 1.03 37 -0.05 1.01 35 0.26 1.00 1.04 0.20 74 0.02 0.92 36 0.07 1.20 37 -0.12 0.94 36 -0.07 0.72 17 0.19 1.13 19 0.15 Variability 75 0.11 1.08 37 0.05 1.16 38 -0.12 0.81 37 -0.10 0.76 18 0.04 1.14 19 0.11 MRT speed Baseline Speed MRT - 0.99 185 0.00 0.87 9 -0.29 0.71 12 -0.11 0.90 16 0.00 0.98 18 0.12 1.02 17 0.06 1.01 38 0.20 1.06 20 -0.18 0.91 20 -0.08 1.15 18 -0.15 0.99 7 0.35 1.41 10 -0.14 - 0.99 185 0.00 0.95 9 0.05 1.42 12 -0.17 0.75 16 -0.32 0.73 18 0.11 0.87 17 0.02 1.07 38 0.16 1.21 20 -0.05 1.14 20 -0.11 0.76 18 0.27 0.73 7 -0.44 1.00 10 -0.01 Accuracy Facial Emotion Recognition - 1.00 373 0.00 1.16 18 -0.40 0.95 18 -0.03 1.14 38 -0.12 0.92 38 -0.05 0.86 38 0.02 0.97 74 -0.06 1.10 37 0.19 0.85 38 0.03 1.05 36 0.06 1.21 19 0.24 0.90 19 0.13 - 1.00 374 0.00 0.98 18 -0.63 0.95 18 -0.29 0.92 38 -0.28 1.13 37 -0.11 1.05 38 0.00 1.04 74 0.14 0.85 37 0.30 0.82 38 0.31 1.03 38 0.02 1.10 19 0.03 0.72 19 -0.17 Visuospatial Verbal Attention 0.80 1.00 349 0.00 1.02 16 -0.56 0.96 17 -0.26 0.93 37 -0.13 0.95 34 -0.19 1.12 37 -0.07 0.99 71 0.05 1.01 34 0.11 1.08 36 0.03 0.81 35 0.14 0.80 16 0.65 1.05 16 0.24 0.28 1.00 374 0.00 0.63 18 -0.19 1.08 18 -0.28 1.08 38 -0.28 0.88 37 -0.10 0.97 38 -0.17 1.01 74 0.01 0.89 37 0.47 1.08 38 0.11 0.94 38 0.04 1.12 19 0.16 1.08 19 0.09 Visuospatial Verbal Working memory Cognitive tasks 1.05 1.00 377 0.00 1.00 18 -0.59 0.97 18 -0.30 1.10 38 -0.35 0.78 38 -0.28 1.02 38 0.13 0.99 75 0.08 1.12 38 0.12 1.01 38 0.04 0.85 38 0.26 0.92 19 0.26 0.79 19 - 1.00 366 0.00 0.81 18 0.00 1.13 17 0.10 1.01 37 -0.11 1.18 37 0.06 0.92 34 0.04 1.01 73 -0.06 1.05 38 -0.10 1.03 38 -0.14 0.93 36 0.06 0.73 19 0.28 1.01 19 0.24 - 1.00 367 0.00 1.13 18 -0.20 1.12 17 0.24 1.16 37 -0.24 0.92 37 0.17 0.99 35 -0.13 0.96 73 -0.01 0.97 38 0.17 0.90 38 -0.07 0.97 36 -0.03 0.94 19 0.06 1.10 19 0.18 Response Accutime racy Visual patterns 0.46 Motor inhibition Block patterns - 1.00 368 0.00 0.83 18 -0.10 0.90 17 -0.19 1.04 38 -0.18 1.05 37 -0.05 0.95 35 0.00 1.04 74 0.00 1.02 37 0.23 1.08 38 0.01 1.07 36 0.05 0.74 19 0.20 0.95 19 -0.04 - 1.00 368 0.00 1.14 18 -0.27 1.24 17 -0.03 1.13 38 0.02 1.01 37 0.16 0.90 35 -0.08 1.04 74 -0.07 0.94 37 0.19 0.91 38 -0.10 0.94 36 0.05 0.89 19 0.06 0.92 19 0.03 Response Accutime racy Cognitive flexibility Note. SWAN=Strengths and Weaknesses of ADHD symptoms and Normal behavior scale. M=mean. SD=standard deviation. N=number of individuals. MRT=mean reaction time. Internalizing and externalizing problems and cognitive tasks were normalized using a van der Waerden transformation and transformed into z-scores to achieve a mean of 0 and a standard deviation of 1 (see row labeled: Total [0-100%]). All measures were corrected for age and sex. The SWAN measure was categorized into quantiles. Quantiles divide a frequency distribution into ordered groups each containing roughly the same number of individuals. 20%, 10% and 5% quantiles were created, which divided the frequency distribution into 5, 10 and 20 ordered groups, respectively. Greatest resolution is provided at the top and bottom extremes of the SWAN distribution (for which 5% quantiles are shown), and least resolution at average levels of the SWAN where data points were the most dense (for which the 20% quantile is shown). Mean behavior problem and cognitive scores were then tabulated for each ADHD quantile. Through comparisons of quantiles at the top and bottom extremes it was possible to examine whether the non-symptomatic ends of the ADHD distribution were on average linked to better behavioral and cognitive outcomes than the symptomatic ends. The 95-100% quantile included the children with the most symptomatic scores on the SWAN, the 0-5% quantile children with the least symptomatic scores. SMD (standardized mean difference) refers to a comparison between children at the extremes comparing the 0-5% versus 95-100% quantile, expressed as a difference in means based on a standard deviation of 1. 1.95 1.00 SD SMD 378 0.00 N 0.95 SD M 18 (0-100%) 1.29 0.99 SD N 19 M 0.80 0.89 SD N 38 M 0.45 0.84 SD N 38 M 0.02 0.82 SD N 38 M -0.13 0.91 SD N 75 M -0.11 0.91 SD N 38 M -0.07 0.85 SD N 38 M -0.17 1.04 SD N 38 M -0.39 0.92 SD N 19 M -0.35 1.01 SD N 19 N M -0.66 M Total 95-100% 90-95% 80-90% 70-80% 60-70% 40-60% 30-40% 20-30% 10-20% 5-10% 0-5% ADHD quantiles EmoOppo- tional Perfect- Psychositional lability Anxiety ionism somatic Internalizing and externalizing problems Supplement 2.3 Mean internalizing and externalizing behavior and mean cognitive task scores per ADHD (SWAN) trait quantile Chapter 2 57 Chapter 2 Supplement 2.4 Scatterplots of linear and quadratic relations between the ASD (AQ) trait and visuo-spatial working memory Supplement 2.5 Scatterplots of linear and quadratic relations between the ADHD (SWAN) trait and oppositional behavior, emotional lability and perfectionism a. a. b. b. Note. Visuo-spatial working memory scores are normalized (van der Waerden transformed), age and sex corrected z-scores. Results for visuo-spatial working memory are shown separately for boys and girls, as the quadratic relation between the AQ and visuo-spatial working memory was only significant in girls, but not boys. c. Note. Oppositional behavior, emotional lability and perfectionism scores are normalized (van der Waerden transformed), age and sex are corrected z-scores. 58 59 Are autism spectrum disorder and attention-deficit/hyperactivity disorder different manifestations of one overarching disorder? Cognitive and symptom evidence from a clinic and population-based sample Jolanda M. J. van der Meer, Anoek M. Oerlemans, Daphne J. van Steijn, Martijn G. A. Lappenschaar, Leo M. J. de Sonneville, Jan K. Buitelaar, Nanda N. J. Rommelse Journal of the American Academy of Child and Adolescent Psychiatry, 2012; 51(11), 1160-1172. 60 Abstract Autism Spectrum Disorder (ASD) and Attention-Deficit/Hyperactivity Disorder (ADHD) frequently co-occur. Given the heterogeneity of both disorders, several more homogeneous ASD-ADHD comorbidity subgroups may exist. The current study examined if such subgroups exist and whether their overlap or distinctiveness in associated comorbid symptoms and cognitive profiles gave support for a gradient overarching disorder hypothesis or a separate disorders hypothesis. Therefore, latent class analyses (LCA) were performed on Social Communication Questionnaire (SCQ) and Conners’ Parent Rating Scale (CPRSR:L) data of 644 children (5 -17 years). Classes were compared for comorbid symptoms and cognitive profiles of motor speed and variability, executive functioning, attention, emotion recognition and visual spatial pattern recognition. LCA revealed five classes: two without behavioral problems, one with only ADHDbehavior, and two with both clinical symptom levels of ASD and ADHD, but with one domain more prominent than the other (ADHD[+ASD] and ASD[+ADHD]). In accordance with the gradient overarching disorder hypothesis were the presence of an ADHD class without ASD symptoms, and the absence of an ASD class without ADHD symptoms, as well as cognitive functioning of the simple ADHDclass being less impaired than that of both comorbid classes. In conflict with this hypothesis was that there was some specificity of cognitive deficits across classes. The overlapping cognitive deficits may be used to further unravel the shared etiological underpinnings of ASD and ADHD, while the non-overlapping deficits may indicate why some children develop ADHD despite their enhanced risk for ASD. The two subtypes of children with both ASD and ADHD behavior will most likely benefit from different clinical approaches. 63 Chapter 3 Autism Spectrum Disorder (ASD) and Attention-Deficit/Hyperactivity Disorder This is a significant step forward and will likely boost research on the shared (ADHD) are both severely impairing, highly heritable neurodevelopmental disorders. and specific pathways related to comorbid ASD and ADHD. However, given that ASD is characterized by impaired social interaction skills and communication, both separate disorders are rather heterogeneous in symptom presentation, as well as restricted and repetitive behavior and interests, whereas ADHD is associated cognitive deficits and underlying etiological factors (Verté et al., 2006), characterized by severe inattention, hyperactivity and impulsivity (American it is likely that children with a comorbid diagnosis of ASD and ADHD form a very Psychiatric Association, 2000). Even though the DSM-IV diagnostic criteria of both heterogeneous group as well. This might hinder both clinical interventions as well disorders appear to show little overlap, both disorders frequently co-occur. The as scientific studies on the etiology of comorbid ASD and ADHD (Rommelse et majority of comorbidity estimates reported for ADHD in ASD fall within the 30 to al., 2010). 80% range, whereas the presence of ASD is estimated in 20 to 50% of the patients with ADHD (Ames & White, 2011; Leyfer et al., 2006; Ronald et al., 2008). Features ADHD may be achieved using an empirical bottom-up approach, such as latent such as poor social skills, language delay, sensory over-responsivity, attention class analyses (LCA). This statistical method allows classification into classes on problems, oppositional defiant behavior and emotion regulation problems are the basis of type and severity of ASD and ADHD symptoms, composing mutually frequently disclosed in both ASD and ADHD (Gadow et al., 2009; Mulligan et exclusive classes (Mc Cutcheon, 1987). By using quantitative symptom measures al., 2009; Rommelse et al., 2011). Furthermore, there is an increasing number reflecting the nature and severity of ASD and ADHD symptoms, justice is done to of studies documenting on various patterns of association between ASD and the continuously distributed nature of symptoms of both ASD and ADHD within ADHD (St. Pourcain et al., 2011) and an overlap between ASD and ADHD with the population (Kim et al., 2011; Levy, Hay, McStephen, Wood & Waldman, 1997). respect to genetic factors (Ronald et al., 2008) and cognitive functions linked This approach has already been successfully applied in the separate research to a familial vulnerability for ASD and ADHD (intermediate phenotypes), such fields of ASD and ADHD, and has with greater precision disclosed clinically as executive functioning, motor speed and variability, emotion recognition and relevant, more homogeneous subtypes of both disorders (Acosta et al., 2008; visual spatial functioning (Booth & Happé, 2010; Corbett, Constantine, Hendren, Constantino, 2011; Volk, Todorov, Hay & Todd, 2009). Of great interest is the Rocke, Ozonoff, 2009; Fine, Semrud-Clikeman, Butcher & Walkowiak, 2008). study of Reiersen, Constantino, Volk, & Todd (2007) that examined ASD-traits in These findings suggest that there are shared etiological pathways for ASD and several population-derived ADHD latent classes. Almost all classes (except one ADHD and patients with one of both disorders should be routinely checked for the class that merely showed hyperactive symptoms) showed more ASD symptoms presence of the other disorder. Furthermore, as the clinical presentation of both than the class without ADHD symptoms. However, amongst the various ADHD disorders is strongly influenced by age, this monitoring should occur on a regular classes, severity of ASD traits clearly differed, with the class having the most basis (St. Pourcain et al., 2011). severe ADHD symptoms also displaying the most severe ASD symptoms. Although the current psychiatric classification scheme prevents a Using DSM-IV derived ADHD subtypes, less distinctions between the subtypes diagnosis of ADHD in the context of ASD (American Psychiatric Association, 2000), were present, with no significant difference between the ADHD predominantly based on the assumption that ASD mimics or even causes symptoms of ADHD, inattentive subtype and the ADHD predominantly hyperactive-impulsive subtype. a comorbid diagnosis of ASD and ADHD can be made in the upcoming DSM-5. This study shows that empirically defined symptom classes may be more useful 64 Identification of more homogeneous subgroups of patients with ASD and 65 Chapter 3 in examining comorbidity patterns across the continuum of symptom severity. studies used DSM-IV defined (hence heterogeneous) groups of patients and However, since no previous study has used both ASD and ADHD symptom ADHD was not always assessed in children with ASD (Rommelse et al., 2011), measures in LCA, it is as yet unknown to what extent mutually exclusive ASD- group comparisons may have been significantly obscured. ADHD classes can be identified. By examining the overlap and distinctiveness of associated traits (such homogeneous classes exist and whether their overlap or distinctiveness in as comorbidity patterns and cognitive problems) between the various ASD-ADHD associated traits (comorbid symptoms, such as oppositional behavior, emotional classes, several hypotheses can be tested. First of all, it has recently been postulated lability, anxiety, perfectionism and psychosomatic complaints, and cognitive that ASD and ADHD are different manifestations of one overarching disorder (H1: profiles) gave support for the (gradient) overarching disorder hypotheses or for overarching disorder hypothesis) (Rommelse et al., 2011). If true, symptomatic the (partly) separate disorders hypothesis. Continuous ASD and ADHD symptom expression can be regarded as ‘noise’ and classes (if at all identified) will not questionnaire data were available for 644 children aged between 5 and 17 years show distinctiveness in associated traits. Second, a variant of this hypothesis from a clinic and population-based sample. Given the distinct developmental states that ADHD may best be seen as a milder, less severe subtype within the characteristics of both disorders, the influence of age was taken into account. A ASD spectrum (H2: gradient overarching disorder hypothesis) (Rommelse et al., large variety of cognitive domains was assessed, most robustly associated with 2011). LCA will then identify at least one ADHD class without ASD symptoms, but ASD (e.g. identification of facial emotions, cognitive flexibility and detail-focused no ASD class without ADHD symptoms, and all classes will show rather similar visual spatial processing) or ADHD (e.g. motor speed and variability, inhibition, associated traits with lowest severity in the ADHD class without ASD symptoms verbal and visual attention, and verbal and visual-spatial working memory), as and highest severity in the ASD with ADHD class. Alternatively, ASD and ADHD documented in previous studies (Booth & Happé, 2010; Corbett et al., 2009; do not constitute different expressions of one overarching disorder. In this case, Fine et al., 2008; Rommelse et al., 2011). LCA were used to identify phenotypical the LCA will identify at least some classes with pure ADHD or ASD symptoms. homogeneous classes and it was examined to what extent overlap and specificity Further, the classes will be more different than similar in terms of associated traits in comorbid symptoms and cognitive profiles existed between the classes. Therefore, the current study set out to examine if different ASD-ADHD (H0: distinct disorders hypothesis). Some previous studies using DSM-IV defined subgroups of patients support the gradient overarching disorder hypothesis, Methods with ASD children having more severe, but similar type of cognitive problems Participants compared to children with ADHD (Gadow, DeVincent & Pomeroy, 2006; Gadow et al., 2009; Geurts, Verté, Oosterlaan, Roeyers & Sergeant, 2004; Holtmann, Bolte, & Poustka, 2007; Nydén et al., 2010). In contrast, there is also evidence for the (partly) distinct disorder hypothesis, with ASD children having qualitative different cognitive problems than children with ADHD (Booth, Charlton, Hughes, Happé, 2003; Buhler, Bachmann, Goyert, Heinzel-Gutenbrunner, & Kamp-Becker, 2011; Sinzig, Morsch, Bruning, Schmidt, & Lehmkuhl, 2008). However, because these 66 The study has been approved by the Committee on Research involving Human Subjects (CMO) and participants were enrolled between January 2009 and July 2011. Eligible participants were 360 children from a random population cohort study (Schoolkids Project Interrelating DNA and Endophenotype Research; SPIDER) and 254 children from a clinic-based ASD-ADHD genetic study (Biological Origins of Autism; BOA). The BOA cohort consisted of siblings, including DSM-IV based ASD, ADHD and ASD+ADHD cases and non-affected 67 Chapter 3 siblings (for a full description, see Box 1.2 regarding the study samples, or see van Steijn et al., 2012). All children were between 5 and 17 years of age, of Caucasian descent and had an estimated total IQ of at least 70 on the Wechsler Preschool and Primary Scale of Intelligence (WPPSI), Wechsler Intelligence Scale (WISC-III) or Wechsler Adult Intelligence Scale (WAIS-III) (Wechsler 1989; 2000; 2002). Exclusion criteria were epilepsy, known genetic or chromosomal disorders (such as Down syndrome), brain damage, and problems with vision or hearing. After complete description of the study to the parents, written informed consent was obtained. Measures Data analyses In order to identify homogeneous symptom classes, LCA were used on the subscale outcomes of the SCQ (social interaction, communication and stereotypic behavior) and the subscale scores of the following ten scales of the CPRS-R:L: inattention, restlessness, cognitive problems, hyperactivity, oppositional behavior, emotional lability, anxiety, perfectionism and psychosomatic complaints. Subscales which represented the total of other subscales (Global Index Total) and subscales which restructured the items by DSM-criteria (DSM-IV inattention, DSM-IV hyperactivity-impulsivity and DSM-IV total) were excluded to prevent overrepresentation of these items. The LCA were carried out using Mplus version 6.11 (Muthén & Muthén, 2006). Both the probability for a child to belong to each ASD and ADHD symptom measures of the classes and the conditional probabilities for children in a particular class to ASD and ADHD symptom measures (parent reports) were taken from the Social show specific behavior were estimated. Children could only be admitted to one Communication Questionnaire (SCQ, Lifetime version) and the Conners’ Parent of the classes. Class differences with respect to sex, age and IQ were analyzed Rating Scale-Revised: Long version (CPRS-R:L). The SCQ and the CPRS-R:L are to check for possible confounders. If differences were detected, only age was both validated instruments for screening ASD and ADHD (Conners et al., 1998a; implemented as covariate in further analyses, since both IQ and sex are inherently Rutter, Bailey, Berument, Lecouteur, Lord, Pickles, 2003). confounded with ASD and ADHD, and therefore could not be separated from the effects of the disorders (Dennis et al., 2007). Class by age interaction effects Procedure were examined and, if significant, post-hoc analyzed and reported. If non- The tasks described and presented in Supplement 3.1 were part of the broader significant, these interactions were dropped from the model. Next, mean factor neuropsychological assessment batteries used in the BOA and SPIDER projects. sum scores of all behavioral domains were computed, and presented in a line Children completed the battery in approximately two hours and the order of task chart, so that quantitative differences between classes could be examined. Size administration was counterbalanced. Due to time constraints, not all tasks were and significance of differences were determined with a MANOVA, after which the administered to all children. Full-Scale IQ was prorated by four subtests of the Bonferroni correction for multiple testing was used for all post-hoc comparisons. WPPSI, WISC-III or WAIS-III (Wechsler 1989; 2000; 2002); Block Design, Picture Completion, Similarities and either Vocabulary (BOA) or Arithmetic (SPIDER). All variables were successfully normalized and standardized into z-scores by These subtests are known to correlate between .90 and .95 with Full-Scale IQ applying a Van der Waerden transformation (SPSS version 18). Effect sizes (Groth-Marnat 1997, Kaufman, Balgopal, Kaufman, McLean, 1994). Parents were were defined in terms of percentage of variance explained (ηp2). Small, medium invited to fill in several questionnaires concerning their youngster’s behavior. and large effects were defined in variances of 0.01, 0.06 and 0.14 respectively Secondly, the underlying cognitive profiles of the classes were examined. (Cohen, 1988). Some of the outcome measures were mirrored, so that the 68 69 Chapter 3 scores of all variables would imply the same: a higher z-score was indicative of 0.003). The behavioral profiles of the classes are presented in Figure 3.1. For a better performance. The classes were compared for each domain separately the sake of clarity, the classes were labeled. Classes 1 and 2 could be viewed as using ANCOVA’s with class-membership as a fixed factor, age as a covariate and norm groups, showing hardly any problems on the separate behavioral domains. speed, accuracy or variability separately for each domain as dependent variable. Therefore, these classes were labeled ‘Normal’ (n = 268 and n = 150). Next, Correction for multiple comparisons was applied according to the False Discovery class 3 was best referred to as ‘ADHD’, with only ADHD-behavioral problems Rate (FDR) controlling procedure to the analyses with a q-value setting of 0.05 (n = 109). Here, both DSM-IV-oriented CPRS-subscales for ADHD (Inattentive (Benjamini & Hochberg, 1995). Only the effects that remained significant after the and Hyperactive-Impulsive behavior) were above clinical cut-off, whereas the FDR-correction were reported. SCQ total score was substantially below cut-off (see Table 3.1). Classes 4 and 5 consisted of children who scored high on both ADHD (CPRS) as well as ASD- Results symptoms (SCQ). In class 4, the ADHD-symptoms were more prominent than Identifying Homogeneous Symptom Classes the ASD symptoms, with CPRS-subscales for ADHD-symptoms substantially The LCA were based on fit and accuracy measures and revealed a solution with five classes (Nylund et al., 2007). Five classes had the best fitting BIC and SSA BIC values, entropy (.914), and p-values on the Vuong-Lo-Mendell-Rubin Likelihood Ratio Test and Lo-Mendell-Rubin Adjusted Likelihood Ratio Test (both p’s = Figure 3.1 Class scores on SCQ (left) and CPRS (middle and right) subscales Class 5 ASD(+ADHD) 9.0 % 1 0,9 Mean sum scores 0,8 Class 4: ADHD(+ASD) 9.2 % 0,7 0,6 0,5 Class 3: ADHD 16.9 % 0,4 0,3 0,2 Class 2: Normal 23.2 % 0,1 om m C So ci al in te ra ct io n un ic at io St n er eo So t y ci pi al c pr ob le m s In at te nt R io es n tle ss ne ss C pr ogn ob it i H lem ve yp er s ac tiv O ity pp os Em iti o ot na io l na ll ab ilit y An x ie Pe ty rfe Ps ctio yc nis co ho m m so pl m ai a nt tic s 0 Class 1: Normal 41.6 % 0,3 70 ores 0,2 0,1 (see Table 3.1). Therefore, class 4 was described as ‘ADHD(+ASD)’ (n = 58). In contrast, in class 5 ASD-symptoms were more prominent than ADHD symptoms, with the SCQ total score being substantially above the clinical cut-off, the CPRSHyperactive-Impulsive subscale just above the cut-off and the CPRS-Inattention subscale slightly below cut-off (see Table 3.1). Therefore, class 5 was described as ‘ASD(+ADHD)’ (n = 59). No class with only ASD-behavior was identified. A MANOVA using class as fixed factor and all behavioral domains as dependent variables revealed that, as expected, the five classes differed significantly on all behavioral subscales (all p < .001). Next, all classes were reciprocally compared on the separate behavioral domains using Bonferroni corrected posthoc comparisons. Only 14 out of 130 comparisons did not reach significance. Roughly, the non-significant differences were between the ADHD-only class and the ASD(+ADHD) class on the ADHD-behavioral domains (in the middle of Figure 3.1), or on the comorbid behavioral domains, such as anxiety and perfectionism (on the right side of Figure 3.1). When corrected for the influence of age, no behavioral domain Note. A higher mean factor sum score indicated that children in that class lacked more competences or showed more problems on the specific domain. SCQ = Social Communication Questionnaire, CPRS = Conners’ Parent Rating Scale, ASD = subscales 0,5 reflecting ASD-symptoms, ADHD = subscales reflecting ADHD-symptoms, Class 1: LL (10.1%) 0,4 above the clinical cut-off, and a SCQ-total score slightly above the clinical cut-off changes in differences between the classes were found. Comparisons between the two Normal classes (classes 1 and 2) were considered theoretically irrelevant. n = 38 Class 3: MM (54.2%) n = 203 71 Chapter 3 Hence, these classes were considered one class in further analyses. To check whether this would affect further analyses in any possible way, both classes were Table 3.1 Demographic characteristics of the children in the distinct classes Normal ADHDa ADHD (+ASD)a compared on demographic characteristics as well as on all cognitive outcome measures. None of those comparisons reached significance (all p’s >.05); classes 1 and 2 clearly only differed on a behavioral level, most likely ranging from n=418 Age in years M 9.5 n=109 SD M 2.4 9.9 n=59 SD 2.8 M 11.2 ASD (+ADHD)a n=58 SD 3.3 M SD 11.5 2.7 Normal = ADHD < ADHD(+ASD) = ASD(+ADHD) Normal = ADHD < ADHD(+ASD) = ASD(+ADHD) normal up to super normal behavior, but without meaningful cognitive differences. The distribution of all children across the distinct classes, as well as the age, sex, Contrasts based on p-values of .05 % Male 45.7 66.1 81.4 86.2 % Population basedb 71.4 56.4 17.2 5.8 Estimated full scale IQc 106.2 13.8 104.2 13.9 101.5 15.5 104.2 population and IQ distributions are provided in Table 3.1. 13.9 Normal > ADHD(+ASD) 4.1 4.4 6.9 4.7 16.3 7.2 23 6 Normal < ADHD < ADHD(+ASD) < ASD(+ADHD) T-score CPRSe Inattention 47.3 6 64.7 8.3 73.3 8.8 62.6 8.2 Normal < ADHD = ASD(+ADHD) < ADHD(+ASD) T-score CPRSe HyperactiveImpulsive 48.2 6.7 64.8 9.8 79.8 8.2 66.7 11.2 Normal < ADHD = ASD(+ADHD) < ADHD(+ASD) Total score SCQd Note. a ADHD = class with behavioral problems in ADHD only. ADHD(+ASD) = class with severe ADHD-symptoms, who also show ASD-symptoms. ASD(+ADHD) = class with severe ASD-symptoms, who also show ADHD-symptoms. b Percentage of the class derived from the general population. c Full-scale IQ was estimated by four subtests of the WPPSI, WISCIII or WAIS-III (Wechsler 1989; 2000; 2002): Block Design, Picture Completion, Similarities and either Vocabulary or Arithmetic. These subtests are known to correlate .90 to .95 with Full-scale IQ (Groth-Marnat 1997). d The total score on the SCQ (Social Communication Questionnaire) reflected the total amount of ASD-symptoms. The official cut-off score for probable ASD is 15, and for definite ASD the cut-off is 21. e Subscale scores on the CPRS (Conners’ Parent Rating Scale) subscales reflected the degree of ADHD-related symptoms. The official cut-off for clinically relevant ADHD-symptoms is a subscale score above 63. Cognitive Profiles of the Distinct Classes To test in which cognitive domains the classes overlapped or differed, separate ANCOVA’s were used for each cognitive domain, with age as a covariate. Results are also presented in Figures 3.2a-g. In these figures, Normal refers to the classes with no behavioral problems. ADHD refers to the class with behavioral problems in ADHD only. ADHD(+ASD) refers to the children who most prominently show 72 73 Chapter 3 severe ADHD-symptoms, but who also show ASD-symptoms. ASD(+ADHD) refers to the class with profound ASD-symptoms, but who also show ADHDsymptoms. The means are adjusted for the covariate age and error bars represent 1 standard error. Identification of Facial Emotions Significant medium and small effects of class were found for speed (F (3,379) = 7.52, p < .001, ηp2 = .06) and accuracy (F (3,379) = 5.39, p = .001, ηp2 = .04), respectively. Pairwise comparisons revealed that responses in the ASD(+ADHD) class were significantly slower and less accurate than the Normal Baseline Speed and Variability class (both p’s < .001) and significantly slower than the ADHD-class (p = .001). A significant but small effect of class was found for speed (F (3,611) = 4.91, p The ADHD(+ASD) class was also slower than the Normal class (p = .007), and = .002, ηp2 = .02) and variability (F (3,611) = 3.41, p= .017, ηp2 = .02). Pairwise formed an intermediate class that did not differ from the other classes regarding comparisons indicated that the ASD(+ADHD) class responded significantly slower accuracy. and more variable than the Normal class (p = .001 and p = .008, respectively). The ADHD-only and ADHD(+ASD) groups formed intermediate classes that did not differ from the other classes. b. Figure 3.2b Differences between the classes on measures of identification of facial emotions Speed a. Figure 3.2a Differences between the classes on measures of baseline speed p = .001 0,3 Z-score faster and less variable faster and more accurate 0,7 0,5 0,1 -0,1 p = .007 0,3 p = .001 0,1 -0,1 -0,3 -0,5 p < .001 -0,3 -0,7 -0,5 -0,7 p < .001 0,5 Variability Z-score Speed Accuracy 0,7 p = .008 Normal (n = 405) ADHD (n = 101) Normal (n = 234) ADHD (n = 64) ADHD(+ASD)(n=45) ASD(+ADHD)(n=41) Note. Group differences presented were based on a mean age of 10.2 years. ADHD(+ASD)(n=58) ASD(+ADHD)(n=52) Note. Group differences presented were based on a mean age of 9.9 years. Inhibition and Cognitive Flexibility No significant effect of class was found for speed or accuracy in motor inhibition (F (3,605) = 2.14, p = .09, ηp2 = .01 and F (3,605) = 0.92, p = .432, ηp2 = .005, respectively) or in cognitive flexibility (F (3,605) = 1.19, p = .31, ηp2 = .006 and F (3,605) = 0.60, p = .614, ηp2 = .003, respectively). 74 75 Chapter 3 c. Figures 3.2c and 3.2d Differences between the classes on measures of motor inhibition and cognitive flexibility Speed Accuracy Z-score 0,5 class (p = .003). Furthermore, all classes performed worse than the Normal 0,3 class in the verbal attention task (p = .007 for the ADHD-class, p < .001 for the comorbid classes). Other group differences did not reach significance. A small 0,1 but significant class by age interaction effect was found in the forward trials for -0,1 visuo-spatial attention (F (3,607) = 4.37, p =.005, ηp2 = .02). Post-hoc analyses -0,3 indicated that the age-effect was stronger in the Normal and ADHD-only classes compared to both comorbid classes, resulting in larger class-differences in older -0,7 than younger children (F (1,449) = 8.50, p =.004, ηp2 = .02 for the Normal class Normal (n = 403) ADHD (n = 100) Speed ADHD(+ASD)(n=56) ASD(+ADHD)(n=51) Normal class compared to ADHD(+ASD) class, (F (1,149) = 7.64, p =.006, ηp2 Accuracy = .05 for the ADHD class compared to the ASD(+ADHD) class, and (F (1,155) = 0,7 4.09, p =.045, ηp2 = .03 for the ADHD-only class compared to the ADHD(+ASD) 0,5 class). e. 0,3 Figure 3.2e Differences between the classes on measures of attention Visuo-Spatial Attention 0,1 Verbal Attention 0,7 -0,1 p = .003 0,5 -0,3 -0,5 -0,7 Normal (n=403) ADHD (n=100) ADHD(+ASD) (n=56) ASD(+ADHD) (n=51) Note. Group differences presented were based on a mean age of 9.9 years. Visuo-Spatial and Verbal Attention p < .001 0,3 Z-score Z-score faster and more accurate compared to ASD(+ADHD) class, (F (1,455) = 5.34, p =.021, ηp2 = .01 for the more accurate responses faster and more accurate and the ASD(+ADHD) class performed worse in comparison to the Normal -0,5 p = .013 0,1 -0,1 p = .007 -0,3 -0,5 Small significant effects of class were found for accuracy in the forward trials for visuo-spatial attention (F (3,607) = 7.95, p < .001, ηp = .04) and verbal attention 2 76 visuo-spatial task indicated that the ADHD(+ASD) class performed worse than the Normal class and ADHD-only class (p < .001 and p = .013 respectively) 0,7 d. (F (3,623) = 10.04, p < .001, ηp2 = .05), respectively. Pairwise comparisons for the -0,7 p < .001 p < .001 Normal (n=411) ADHD (n=105) ADHD(+ASD) (n=56) ASD(+ADHD) (n=56) Note. Group differences presented were based on a mean age of 9.9 years. 77 Chapter 3 compared to the ADHD-class and F (1,434) = 5.19, p =.023, ηp2 = .01 for the Visuo-Spatial and Verbal Working Memory Small but significant effects of class were found for accuracy in the backward trials for both visuo-spatial working memory (F (3,581) = 5.10, p =.002, ηp2 = .03) and verbal working memory (F (3,623) = 5.68, p = .001, ηp2 = .03). Pairwise comparisons indicated that in the visuo-spatial task, the ADHD(+ASD) class performed worse than the Normal class (p < .001). In the verbal working memory task, the ADHD-class performed worse than the Normal class (p < .001). The ASD(+ADHD) class formed an intermediate group, not differing from the other classes regarding both visuo-spatial and verbal working memory. A small but significant class by age interaction effect was found in the backward trials for visuo-spatial working memory (F (3,581) = 5.37, p =.001, ηp2 = .03). Post-hoc analyses indicated that the effect of age was stronger in the Normal class than in the three clinical classes, with larger class differences in older than younger children (F (1,428) = 12.16, p =.001, ηp2 = .03 for the Normal class compared to the ASD(+ADHD) class, F (1,474) = 4.35, p =.038, ηp2 = .01 for the Normal class f. Figure 3.2f Differences between the classes on measures of working memory Visuospatial Working Memory Verbal Working Memory 0,7 A small but significant effect of class was found for accuracy (F (3,634) = 6.93, p < .001, ηp2 = .03) (also when processing speed or IQ was implemented as a covariate; p < .001 and p = .003, respectively). Pairwise comparisons revealed that the ADHD(+ASD)-class performed significantly worse compared to the Normal class and the ASD(+ADHD) class (both p’s < .001). The ADHD-only class formed an intermediate group, not differing from any of the other classes. The ASD(+ADHD) class showed the highest score (although not significantly different from the normal class), indicating a detail-focused processing style. A small but significant class by age interaction effect was found for accuracy (F (3,631) = 4.83, p =.002, ηp2 = .02). Post-hoc analyses indicated that the effect of age was stronger in the Normal and ADHD-only classes compared to the ADHD(+ASD) class, resulting in larger class-differences in older than younger children (F (1,471) g. Figure 3.2g Differences between the classes on a measure of visual spatial pattern recognition Detail-focused cognitive style 0,7 0,5 0,1 0,3 -0,1 -0,3 -0,5 -0,7 p < .001 Normal (n=411) ADHD (n=105) ADHD(+ASD) (n=56) ASD(+ADHD) (n=56) Note. Group differences presented were based on a mean age of 9.9 years. Z-score 0,3 more accurate responses Z-score Visual spatial pattern recognition p < .001 0,5 more accurate responses Normal class compared to the ADHD(+ASD) class). p < .001 p < .001 0,1 -0,1 -0,3 -0,5 -0,7 Normal (n=416) ADHD (n=108) ADHD(+ASD) (n=59) ASD(+ADHD) (n=56) Note. Group differences presented were based on a mean age of 9.9 years. 78 79 Chapter 3 = 14.84, p <.001, ηp2 = .03 for the Normal class compared to the ADHD(+ASD) emotional lability symptoms, were significantly lower in the ASD(+ADHD) class class, and F (1,163 = 5.27, p =.023, ηp2 = .03 for the ADHD-only class compared compared to the ADHD(+ASD) class. However, ADHD symptoms may actually to the ADHD(+ASD) class). lead to somewhat better ratings on social interaction in children with ASD, partly because of the increased talkativeness seen with ADHD. Furthermore, some forms Discussion of oppositional behavior may be more closely related to ADHD, whereas others This study aimed at examining if different ASD-ADHD symptom classes exist may be more related to resistance to change (typical of ASD), so the degree of and whether their overlap or distinctiveness in associated traits (comorbid ADHD and ASD symptoms may affect which aspects of oppositional behavior symptoms and cognitive functions) gave support for the (gradient) overarching are most likely to be endorsed. Due to poor communication skills, symptoms disorder hypothesis or for the (partly) distinct disorders hypothesis. If the gradient of emotional lability may be more difficult to detect in children with higher ASD overarching disorder hypothesis is accurate, LCA would identify at least one symptoms. Hence, lower ratings on emotional lability in the ASD(+ADHD) group ADHD class with no or only minor ASD symptoms and no ASD class without may be partially explained because emotional lability goes undetected by the ADHD symptoms. This is exactly what we found. Three patient classes could be measures used. Based on the currently defined classes, the ADHD-only group distinguished from two normal classes: One class with ADHD symptoms only, was indeed the mildest affected class within the spectrum, but no such severity one class with clinically high levels of ADHD but also clinically elevated levels of distinction could be made between both comorbid ASD-ADHD classes; they ASD symptoms (ADHD[+ASD]), and one class with clinically high levels of ASD appeared qualitatively different. As yet, no previous studies have examined the symptoms but also clinically elevated levels of ADHD symptoms (ASD[+ADHD]). ADHD symptoms in ASD defined classes to compare our findings with. The As hypothesized, no class with exclusively ASD-symptoms was revealed; all current novel findings suggest that at least two ASD-ADHD comorbidity classes children who expressed ASD-behavior also presented the less severe ‘precursor’ exist that are not merely dissociable on a quantitative basis. of ADHD-behavior. This finding is in accordance with the higher prevalence of children with ASD who meet criteria for ADHD (up to 80% in literature), compared overarching disorder hypothesis, but on the other hand possible support for the to the ASD rate in children with ADHD (up to 50% in both the literature as well distinct disorder hypothesis emerged as well. That is, the cognitive functioning as our sample) (Ames & White, 2011; Leyfer et al., 2006). These data are also in in the simple ADHD-class could overall be considered at an intermediate level, line with previous studies addressing ASD-symptoms in children with ADHD, in performing somewhat below the normal class and better than the two comorbid- which LCA revealed that the most severe ADHD classes were also the classes classes. This was best visible in the domains of motor slowness and variability, with the most severe ASD-symptoms. However, note that these studies examined visuo-spatial and verbal attention and emotion recognition. However, qualitative ASD symptoms in ADHD defined classes, thereby excluding the possibility of differences were also clearly observed, which should perhaps not be too surprising finding ASD classes without ADHD symptoms (Mulligan et al., 2009; Reiersen et since ADHD and ASD are partly defined by specific cognitive problems (i.e. al., 2007). inattentiveness versus detail-focused processing, respectively). Working memory It may seem in contrast to the (gradient) overarching disorder hypothesis deficits were significantly more pronounced in both primarily ADHD classes that the level of ADHD symptoms, as well as the level of oppositional and compared to the primarily ASD class and a detail-focused cognitive style in visual 80 The cognitive profiles on the one hand further supported for the gradient 81 Chapter 3 pattern recognition (reflected in normal to superior block design performance) severity of the deficit in social cognition was most pronounced in the ASD(+ADHD) appeared to be present in the ASD(+ADHD) class only. This apparent double class, followed by the ADHD(+ASD) class, but was not present in the ADHD-only dissociation between both comorbid classes suggests that dysfunctions in the class. Previous studies documenting on social cognition deficits in ADHD may information processing style found in the ASD(+ADHD) class cannot merely be therefore possibly be explained by elevated levels of ASD symptoms in these seen as a the sum of dysfunctions in the ADHD-only and the ADHD(+ASD) classes. ADHD-patients (Nijmeijer et al., 2008; Cadesky, Mota & Schachar, 2000; Kats- Apparently, different ASD-ADHD comorbid subtypes exist, with overlap but also Gold, Besser & Priel, 2007). Similarly, a detail-focused (i.e. local) processing style, qualitative difference in cognitive deficits. Similarly, recent studies comparing reflected by a normal to superior block pattern performance, was only apparent in people with ASD, ADHD or ASD+ADHD to controls reported evidence for both the ASD(+ADHD) class and not in both primarily ADHD classes, which performed specific as well as overarching deficits (Nydén et al., 2010; Sinzig et al., 2008). In intermediate or more poorly than the Normal and ASD(+ADHD) classes. This any case, our cognitive findings are in contrast to the hypothesis that ASD and normal to superior performance clearly stands out against the background of ADHD are interchangable manifestations of the same overarching disorder and poorer performance of this class on almost all other domains measured, and symptomatic expression can be regarded as ‘noise’, in which case no cognitive is in line with previous studies (Nydén et al., 2010; Happé 1999). Importantly, differences between the classes were predicted. The overlapping cognitive this feature was absent in the ADHD(+ASD) class, suggesting the origins of the deficits may be used to further unravel the shared etiological underpinnings of ASD symptoms in this class are at least partially different from the origins of the ASD and ADHD, whereas the non-overlapping deficits may indicate why some ASD symptoms in the ASD(+ADHD) class. Conversely, working memory deficits children develop ADHD despite their enhanced risk for ASD and vice versa. appeared mainly to be related to ADHD symptoms, being impaired in both ADHD When specifically viewed at the individual cognitive domains, it is classes but not to that degree in the ASD(+ADHD) class. However, given that remarkable that response slowness and variability were most pronounced in there was no class with ASD symptoms only, it remains uncertain whether ASD the ASD(+ADHD) class and not observed in the ADHD-only class. Increased symptoms are truly unrelated to working memory deficits. In any case, the strong response variability is considered one of the most robust and best replicated relation between ADHD symptoms and working memory deficits concurs with cognitive features of ADHD as defined by the DSM nomenclature (Johnson et previous studies (Nydén et al., 2010; Sinzig et al., 2008) and suggests working al., 2007; Frazier-Wood et al., 2012; Klein, Wendling, Huettner, Ruder & Peper, memory performance may shed more light on the causal pathways for ADHD. 2006). Current findings based on latent classes suggest that response variability None of the classes showed problems in inhibition or cognitive flexibility, which in ADHD actually reenacts on the presence of comorbid symptoms such as ASD- was probably due to the predictable nature of the task, in which children always symptoms, as has been described previously (Geurts et al., 2008). Noteworthy, had to respond. Previously, this same task also did not differentiate between verbal and visual attention were affected in all clinical classes. This finding ADHD and controls (Rommelse et al., 2007), suggesting a more unpredictable may imply that a dysfunction in attention is indicative for neurodevelopmental nature of inhibitory control and cognitive flexibility may be more applicable in disorders in general, unable to differentiate between ASD and ADHD, as has been distinguishing patients from normally developing children (Bekker et al., 2005). reported before (Ames & White, 2011). In contrast, impaired social cognition is a Finally, class differences between the Normal class and the (severely) affected substantially affirmed prime deficit specific for ASD (Loveland et al., 1997). The classes on visuo-spatial attention, visuo-spatial working memory and visual 82 83 Chapter 3 pattern recognition were larger among older compared to younger children. This comorbid ASD-ADHD classes may have more persistent ADHD symptoms (St. may suggest that, within the limitations of this cross-sectional design, children in Pourcain et al., 2011), thus setting new targets for a longitudinal research design. the Normal class improve certain skills more during maturation than children in the (severely) affected classes. This however needs to be further analyzed with the help of a longitudinal research design. Some limitations of this study warrant consideration. First, boys were overrepresented in the three affected classes, whereas they were underrepresented in the Normal class. This was likely due to the fact that ASD and ADHD are more frequently diagnosed in boys than in girls (Kramer, Krueger, Hicks & 2008). However, since this overrepresentation was present in all patient classes, this did not affect comparisons between those classes. Second, questionnaires were used to collect information on behavioral problems. In comparison with clinical interviews, questionnaires tend to overestimate the degree of comorbidity, as questionnaires do not allow for further explanation of questions. Interviews may improve the correct interpretation of questions, and more precisely distinguish normal-range behavior from clinical-range behavior (Tourangeau, Rips & Rasinki, 2000). However, the possible comorbidity-overestimation cannot account for the cognitive differences in the distinct classes, nor can it explain the presence of an ADHD-only class and the absence of an ASD-only class. Third, the nature of our samples might have prevented us from detecting a ‘pure ASD’ class. ASD without ADHD symptoms might be underrepresented in clinical samples and also be relatively rare in population samples. Therefore, very large population-based samples might be best to examine this issue further. Future studies in a purely population-based sample may wish to include both cognitive and symptom measures in tandem in latent class analyses or related statistical approaches, to further increase the etiological homogeneity of the distinct classes. Fourth, children in the currently defined ADHD-only class were younger and also mildest affected within the spectrum. It is of great interest whether (a majority of) these children have a childhood-limited form of ADHD, while the children in both 84 85 Chapter 3 Supplemental Material Identification of Facial Emotions Supplement 3.1 This task was used to measure the capacity to quickly and accurately identify Measures facial emotional expressions (de Sonneville, 1999) Four blocks represented different target emotions: happiness, sadness, anger or anxiety. Children had to Four of the tasks described were selected from the Amsterdam Neuropsychological judge whether or not a face expressed the specified target emotion by pressing Tasks (ANT) program (de Sonneville, 1999). Each computer task contained an instruction trial wherein the examiner provided a typical item of the task, and a separate practice session. Test–retest reliability and validity of the computerized ANT-tasks are satisfactory and have been described and illustrated elsewhere (de Sonneville, 2005). b. a yes/no-button as quickly and accurately as possible. The order of the targeted emotions was randomly assigned. Dependent measures were mean reaction time (in ms) and accuracy for all targeted emotions together. Facial Emotion Recognition Baseline Speed and Variability This task was used to measure the speed and variability of motor output, comparable to a simple reaction time task (de Sonneville, 1999). When a fixation cross in the centre of a computer screen changed into a white square, children pressed a key as quickly as possible. In order to prevent anticipation strategies, the time interval between a response and the emergence of the next square varied randomly between 500 and 2500 ms. Dependent measures were response speed (mean reaction time in ms) and variability (SD of reaction time in ms). Inhibition and Cognitive Flexibility This task consisted of three blocks in which the first block measured baseline speed and accuracy (de Sonneville, 1999). Children had to press a key as soon Baseline Speed and Variability as they noticed a green circle on the left or the right of a fixation cross. They were instructed to press a key on the same side as the stimulus was presented. In the second block, the circles were colored red, and children had to press a key on the opposite side. Motor inhibition was calculated as the difference in percentage of errors or in mean reaction time between blocks one and two. Finally in the third block, both green and red circles appeared in a random order, and both same- Fixation Signal side (compatible) and opposite side (incompatible) responses were required. This was hypothesized to demand for higher levels of cognitive flexibility (Los, 1996). Cognitive flexibility was calculated as the difference in percentage of errors or mean reaction time between block one and the compatible trials of block three. 86 87 Chapter 3 Inhibition and Cognitive Flexibility: compatible and incompatible trials Left compatible Right compatible Left incompatible Visuo-Spatial Attention and Working Memory Right incompatible Visuo-Spatial and Verbal Attention The forward parts of both the Visuo-Spatial Attention task (de Sonneville, 1999) and the Digit Span task of the WPPSI, WISC-III or WAIS-III (Wechsler, 1989; 2000; 2002) were used to obtain an indication of visuo-spatial and verbal attention. In the Visuo-Spatial Attention task, stimuli consisted of nine squares, presented in a three by three square. During each trial, a sequence of these squares was pointed at, and the children were instructed to exactly reproduce the sequence. In the Digit Span task, children had to repeat a sequence of verbally presented numbers. In both tasks, the difficulty level increased after each succeeded trial. Dependent variables were the total number of correct sequences in identical order, for both tasks separately. Detail-focused processing style The Block Design task of the WPPSI, WISC-III or WAIS-III (Wechsler, 1989; 2000; 2002) was used to measure detail-focused (local) processing. Children had to copy geometric white-and-red designs using four to nine plastic cubes. All cubes had two completely white, two completely red and two diagonally white-and-red sides. Dependent measure was the score based on the amount of correct and timely completed geometric design. Visuo-Spatial and Verbal Working Memory The backward parts of the Visuo-Spatial Attention task (de Sonneville, 1999) and the Digit Span task (Wechsler, 1989; 2000; 2002) were used to obtain an indication of working memory. Here, children were asked to reproduce the verbal and visuo-spatial sequences (such as described above) in backwards order. Again, the sequence increased if a child reproduced the previous trial successfully. Dependent measures were the total number of correct sequences in backwards order, for both tasks separately. 88 89 How ‘core’ are motor timing difficulties in ADHD? A latent class comparison of pure and comorbid ADHD classes Jolanda M. J. van der Meer, Catharina A. Hartman, Andrieke J. A. M. Thissen, Anoek M. Oerlemans, Marjolein Luman, Jan K. Buitelaar, Nanda N. J. Rommelse Under review 90 Abstract Children with Attention-Deficit/Hyperactivity Disorder (ADHD) often have motor timing difficulties. This study examined whether affected motor timing accuracy and variability are specific for ADHD, or that comorbidity with Autism Spectrum Disorder (ASD) contributes to these motor timing difficulties. Therefore, an 80-trial motor timing task measuring accuracy (μ), variability (σ) and infrequent long response times (τ) in estimating a 1-second interval was administered to 283 children and adolescents (8-17 years) from both a clinic and populationbased sample. They were divided into four latent classes based on the Social Communication Questionnaire (SCQ) and Conners’ Parent Rating Scale (CPRSR:L) data. These classes were: without behavioral problems ‘Normal-class’ (n= 154), with only ADHD symptoms ‘ADHD-class’ (n=49), and two classes with both ASD and ADHD symptoms, but with one domain more prominent than the other; ADHD(+ASD)-class (n= 39) and ASD(+ADHD)-class (n= 41). The pure ADHD-class did not deviate from the Normal class on any of the motor timing measures (mean RTs 916 ms and 925 ms, respectively). The comorbid ADHD(+ASD) and ASD(+ADHD) classes were significantly less accurate (more time underestimations) compared to the Normal class (mean RTs 847 ms and 870 ms, respectively). Variability in motor timing was reduced in the younger children in the ADHD(+ASD) class, which may reflect a tendency to rush a tedious task. Findings suggest that comorbid ASD symptoms contribute to motor timing difficulties in ADHD, as ADHD symptom severity in the pure ADHD-class and the ASD(+ADHD) class was highly similar, with the former class showing no motor timing deficits. 93 Chapter 4 Attention-Deficit/Hyperactivity Disorder (ADHD) is a neurodevelopmental disorder difficulties in motor time processing can be found in ‘pure’ ADHD, or rather are that is typified by developmentally inappropriate degrees of inattention, impulsivity associated with comorbid ASD (Cormier, 2008; Rommelse et al., 2011). Adamo and hyperactivity (American Psychiatric Association, 2000; 2013). Broad patterns and colleagues (2013) compared response time variability in normally developing of neuropsychological impairments have been associated with ADHD, among children, children with ADHD, and children with ASD with and without substantial which deficits in time processing (Castellanos & Tannock, 2002; Falter & Noreika comorbid ADHD symptoms. Findings suggested that both children with ADHD 2011; Noreika, Falter & Rubia, 2013). Falter & Noreika (2011) suggested that and children with ASD and comorbid ADHD had elevated levels of response time deficits in time processing may play an important role in neurodevelopmental variability. In contrast, children with ASD without substantial comorbid ADHD disorders like ADHD by interacting with and modulating primary symptoms. For symptoms did not differ from normally developing children regarding response example, previous studies suggested that difficulties in complex functions such as time variability, suggesting that response time variability is more strongly related attention, language and inhibition are associated with reduced time processing, to ADHD. However, in addition to subtyping along traditional lines of DSM- as these functions are characterized by specific temporal patterns (Rubia et based categories, a comparison of motor timing across homogeneous groups al., 2009; Szelag et al., 2004). Time processing, measured with a motor timing within comorbid ASD-ADHD children may be a powerful method to further paradigm in which the accuracy and variability of motor timing, and infrequent long our understanding of both disorders. Such homogeneous groups based on response times are disentangled by using mu, sigma and tau, may differentially quantitative symptom measures reflect the continuously distributed nature and affect cognitive functions that rely on accurate motor timing (Hervey et al., 2006; severity of ASD and ADHD symptoms across the general population, as shown Leth-Steensen et al., 2000; Thissen et al., (under revision)). Reduced motor by several studies (Constantino, 2011; Fair et al., 2012; Spiker et al., 2002; St time processing has frequently been associated with ADHD, despite systematic Pourcain et al., 2010; 2011). differences across studies, and has shown to be highly heritable, suggestive of an etiological role in ADHD (Andreou et al., 2007; For a review, see Kofler et al., 2013; subgroups of comorbid ASD-ADHD children when studying shared etiological Marx et al., 2010; Noreika et al., 2013; Toplak, Dockstader, & Tannock, 2006).Of pathways in a clinic and population-based sample(van der Meer et al., 2012, see note, abnormalities in motor timing are predominantly related to deficient motor also chapter 3). In that study, classes were derived using a latent class analysis timing processes rather than to general deficient motor functioning in children and (LCA), an empirical method which allows classifications based on the type and adolescents who suffer from ADHD (Rommelse et al., 2008). severity of ASD and ADHD symptoms. We showed that ADHD-symptoms were Despite this compelling evidence for motor timing difficulties in ADHD, present both in the absence and presence of ASD-symptoms. This resulted in reduced time processing has not exclusively been found in ADHD. It has also been a pure ADHD-class that showed no comorbid symptoms of ASD, and an ADHD- observed in other disorders including Autism Spectrum Disorders (ASD)(Allman, class with comorbid ASD (ADHD(+ASD)). Furthermore, ASD-symptoms were DeLeon, & Wearden, 2011; Falter, Noreika, Wearden, & Bailey, 2012; Geurts et reported in the presence of less severe ADHD-symptoms (ASD(+ADHD)), but al., 2008; Martin, Poirier, & Bowler, 2010). Since ADHD is frequently comorbid no class with pure ASD-behavior was identified. The empirical validity of these with ASD (in clinical samples, 20% to 50% of ADHD-patients meet criteria for distinct classes was affirmed by the overlap and distinctiveness of associated ASD; for a review see Rommelse et al. (2011)), it remains to be seen whether comorbidity patterns and cognitive profiles. Classes with children suffering from 94 We previously reported on the advantages of more homogeneous 95 Chapter 4 both types of symptoms were overall cognitively more impaired than children METHODS with only ADHD-symptoms, indicative for an overlapping cognitive background Participants in ASD and ADHD. Importantly, cognitive specificity was found in that the ADHD(+ASD) class showed the more typical ADHD neurocognitive problems (working memory deficits) while the ASD(+ADHD) class showed more typical ASD neurocognitive problems (emotion recognition problems and superior block pattern performance). This cognitive double dissociation between comorbid classes with either more profound ASD or more profound ADHD symptoms can increase our understanding of the distinct etiological pathways for ASD and ADHD (van der Meer et al., 2012). The cognitive domain of time processing is an additional candidate for furthering our understanding of these more homogeneous classes of children affected with pure ADHD or affected with both ASD and ADHD symptomatology. The current study was set out to examine the overlap and distinctiveness in motor timing abilities betweenhomogeneous subgroups with the use of a wellvalidated motor timing paradigm (Rommelse et al., 2008; van Meel et al., 2005). This paradigm measures the accuracy, variability, and infrequent long response times of 1 second interval motor time productions with the use of the parameters mu, sigma and tau, respectively. In sum, the aims were to examine whetherthe 1) accuracy, 2) variability and 3) infrequent long response times differed across the four homogeneous ADHD-ASD symptom classes. Given the previous findings in more homogeneous subgroups (van der Meer et al., 2012), we hypothesized that motor timing is affected (i.e. reduced accuracy, increased variability of motor timing, and increased infrequent long response times) in classes with both ADHD and ASD symptoms, and to a lesser extent, although still different from the Normal class, in the pure ADHD class. The task was randomly assigned to 283 children between 8 and 17 years of age from a population and clinic-based sample. This sample originally consisted of 644 children (van der Meer et al., 2012, see also chapter 3); because of task demands the current task was not administered to the 5, 6 and 7-year olds. 81 children originated from a random population cohort study (Schoolkids Project Interrelating DNA and Endophenotype Research; SPIDER) and 202 children and adolescents from a clinical ASD-ADHD genetic study (Biological Origins of Autism; BOA). The BOA cohort consisted of children with DSM-IV based ASD, ADHD and ASD+ADHD diagnoses and non-affected siblings (for a full description, see Box 1.2 regarding the study samples, or see van Steijn et al., 2012). In the previous study (van der Meer et al., 2012), children were divided into homogeneous symptom classes with the use of a LCA on the raw subscale outcomes of the SCQ (social interaction, communication and stereotypic behavior) and the T-scores of the following ten scales of the CPRS-R:L: social problems, inattention, restlessness, cognitive problems, hyperactivity, oppositional behavior, emotional lability, fear, perfectionism and psychosomatic complaints (for a full description see van der Meer et al., 2012). The raw subscale outcomes of the SCQ and the T-scores of the CPRS used were either unrelated to age (SCQ) or corrected for the influence of age (CPRS), limiting the impact of age on the definition of the latent classes. Five classes had the best fitting BIC and SSA BIC values and entropy (.914), combined with informative class profiles (Nylund et al., 2007). Between class contrasts indicated that the current subsample was comparable to the complete sample regarding ASD symptom severity (all p’s > .06), ADHD symptom severity (all p’s > .08), sex (all p’s > .21), and IQ (all p’s > .21). Consequently, the current sample was older (M (SD) 11.57 (2.5)) than the complete sample (M (SD) 9.5 (2.4)). The distribution of children across the distinct homogeneous symptom classes, as well as the ASD and ADHD symptom severity, age, sex, population and IQ distributions are provided in Table 4.1. These 96 97 Chapter 4 distributions are well in line with the distributions in the complete sample (see also chapter 3). Table 4.1 Demographic characteristics of the children in the distinct classes Normal ADHDa n=49 M SD 11.6 2.5 ADHD (+ASD)a n=39 M SD 12.3 2.7 ASD (+ADHD)a n=41 M SD 12.1 2.5 Age in years n=154 M SD 11.2 2.3 % Male 42.2 69.4 79.5 85.4 %Population basedb Estimated full scale IQc Total score SCQd 39.0 30.6 15.4 0.0 106.7 12.1 104.7 11.8 100.5 12.5 102.2 10.6 Normal > ADHD(+ASD) 4.3 5.4 8.2 5.2 16.4 6.6 22.4 6.0 T-score CPRSe Inattention 48.2 6.5 65.0 7.1 73.3 8.8 62.5 8.6 T-score CPRSe HyperactiveImpulsive T-score CPRS Oppositional behaviore T-score CPRS Cognitive problemse T-score CPRS Anxietye 48.4 7.2 66.3 10.6 79.9 8.8 67.1 12.4 48.3 6.8 57.22 8.3 74.5 9.0 59.4 10.9 49.0 6.5 65.0 8.0 71.8 7.6 59.7 8.3 50.7 11.1 55.4 11.4 71.5 13.6 69.6 13.1 46.6 6.4 49.7 6.4 60.5 11.3 64.8 9.9 T-score CPRS 50.7 Psychosomatic complaintse T-score CPRS 46.1 Emotional labilitye 9.7 54.7 11.8 72.1 15.6 63.0 14.6 7.1 54.5 10.7 71.3 13.9 58.3 10.5 Normal < ADHD < ADHD(+ASD) < ASD(+ADHD) Normal < ADHD = ASD(+ADHD) < ADHD(+ASD) Normal < ADHD = ASD(+ADHD) < ADHD(+ASD) Normal < ADHD = ASD(+ADHD) < ADHD(+ASD) Normal < ASD (+ADHD) < ADHD < ADHD (+ASD) Normal = ADHD < ASD(+ADHD) = ADHD(+ASD) Normal = ADHD < ADHD(+ASD) = ASD(+ADHD) Normal = ADHD < ASD(+ADHD) < ADHD(+ASD) Normal < ADHD = ASD(+ADHD) < ADHD(+ASD) For the sake of clarity, the classes were labeled. Children in class ‘Normal’ showed hardly any problems on ASD-and ADHD behavioral domains (n = 154). Next, class ‘ADHD’ contained children with only ADHD-symptoms (n = 49) without comorbidities. Here, both DSM-oriented CPRS-subscales for ADHD (Inattentive and Hyperactive-Impulsive behavior) were above clinical cutoff, whereas the SCQ total score was substantially below cut-off (see Table 4.1). Children in the class ‘ADHD(+ASD)’ scored above clinical cut-off on both ADHD and ASD-symptoms, with the ADHD-symptoms more prominent than the ASD symptoms (n = 39). Finally, children in the class ‘ASD(+ADHD)’ scored above / at clinical cut-off on both ASD and ADHD-symptoms, with the ASD-symptoms more prominent than the ADHD symptoms (n = 41). No class with only ASD-behavior was identified. All children were of Caucasian descent and had an estimated total IQ of at least 70 on the Wechsler Intelligence Scale (WISC-III) or Wechsler Adult Intelligence Scale (WAIS-III) (Wechsler, 2002; 2000). Exclusion criteria were epilepsy, known genetic or chromosomal disorders (such as Down syndrome), brain damage, and problems with vision or hearing. T-score CPRS Perfectionisme Contrasts based on p-values of .05 Normal < ADHD < ADHD(+ASD) = ASD(+ADHD) Normal < ADHD = ADHD(+ASD) = ASD(+ADHD) Note.a ADHD = class with behavioral problems in ADHD only. ADHD(+ASD) = class with severe ADHD-symptoms, who also show ASD-symptoms. ASD(+ADHD) = class with severe ASD-symptoms, who also show ADHD-symptoms. b Percentage of the class derived from the general population. c Full-scale IQ was estimated by four subtests of the WPPSI, WISC-III or WAIS-III:Block Design, Picture Completion, Similarities and either Vocabulary or Arithmetic (Wechsler, 1989; 2000; 2002). These subtests are known to correlate .90 to .95 with Full-scale IQ(Groth-Marnat, 1997). dThe total score on the SCQ (Social Communication Questionnaire) reflected the total amount of ASD-symptoms. The official cut-off score for probable ASD is 15, and for definite ASD the cut-off is 21.eSubscale scoreson the CPRS (Conners’ Parent Rating Scale) subscales reflected the degree of ADHD-related and comorbid symptoms. The official cut-off for clinically relevantsymptoms on the CPRS is a T-score above 63 98 99 Chapter 4 Measures description of the study to the parents and adolescents, written informed consent Motor Timing Task was obtained. Parents were invited to fill in several questionnaires concerning This 1-second time production task measured the accuracy, variability and infrequent long response times of motor timing (Rommelse et al., 2008; van Meel et al., 2005). Children had to press a button with their preferred index finger when they thought a 1-second time interval had elapsed. The start of the interval was announced by a tone. After the button press, visual feedback concerning the accuracy of the response was presented on screen, indicating whether the their youngster’s behavior. Data analyses Raw responses higher or lower than 4 SD from a subject’s mean, with a minimum response time of 200 ms, were considered outliers and excluded (Leth-Steensen et al., 2000), which was < 0.1% of the data. Slow responses less than 4 SD response was correct, too fast or too slow. A response was regarded as correct above a subject’s mean were not excluded, but represented by τ, the mean of if it fell between the lower and upper boundary set by a dynamic (self-paced) the exponential part of the distribution. Since τ-data were positively skewed, tracking algorithm. Boundaries were set at 500 to 1,500 ms at the beginning normalized z-scores for τ were used in all analyses. These z-scores were obtained of the task (van Meel et al., 2005). If the response fell within these boundaries, the boundaries of the subsequent trial were narrowed by 100 ms. Likewise, the boundaries of the subsequent trial were widened with 100 ms if the response on the previous trial fell outside the boundaries. The practice session consisted of 20 trials, the experimental session consisted of 80 trials. Accuracy of motor timing was represented by μ, the mean of time productions in ms, corrected for the tail of the distribution (infrequent long response times). Variability in motor timing was represented by σ, the variability of time productions in ms corrected for the tail of the distribution (infrequent long response times). Infrequent long response times were represented by τ, the mean of the exponential part of the distribution (Leth-Steensen et al., 2000). Dependent measures μ (mu), σ (sigma) and τ (tau) were calculated with the use of ex-Gaussian analyses performed in MATLAB. by Van der Waerden transformations (SPSS version 20). Effect sizes were defined in terms of percentage of variance explained (ηp2). Small, medium and large effects were defined as explained variances of .01, .06 and .14 respectively (Cohen, 1988). The classes were compared using Repeated Measures ANCOVAs with class-membership as a fixed factor, age, age2 and sex as covariates and, respectively, μ, σ and τ as dependent variables. Age2 was included to adjust for possible nonlinear improvement in task performance across age. Interaction effects were examined and, if nonsignificant, dropped from the model. Correction for multiple comparisons was applied according to the False Discovery Rate (FDR) controlling procedure to the post-hoc analyses with a p-value setting of .05 (Benjamini, 1995). Only the effects that remained significant after FDR-correction were reported. Finally, in light of possible cognitive impairments in unaffected siblings, analyses were repeated excluding unaffected siblings of ASD, ADHD Procedure The task described was part of the broader neuropsychological assessment and ASD+ADHD affected children in the Normal class, to examine a potential influence on the findings. batteries used in the SPIDER and BOA projects. These studies have been approved by the Committee on Research involving Human Subjects (CMO) and children were enrolled between January 2009 and July 2011. After complete 100 101 Chapter 4 RESULTS Variability (σ) Accuracy (μ) No significant class effect was found for the σ (F(3,282) = 0.86, p = .46, ηp2 = A significant class effect, however with small effect size, was found for μ (F (3, .01). A significant class*age interaction effect with a medium effect size, was 282) = 4.20, p = .006, ηp2 = .04). All children seemed to underestimate the 1 found for σ (F (3, 282) = 5.58, p = .001,ηp2 = .06), see also Figure 4.2. Post hoc second interval (see Figure 4.1). Pairwise comparisons indicated that the deviation analyses including two age groups per class indicated that older children in all from the aimed response time (1000 ms) of the ADHD(+ASD) (M = 847 ms) and classes except for the ADHD(+ASD)class showed less variability compared to ASD(+ADHD) (M = 870 ms) classes deviated significantly from that of the Normal their younger counterparts. the only ADHD-class did not differ from that of the other classes (M = 916 ms). No significant class*age interaction effect was found for μ (F(3,282) = 1.71, p = .16, ηp2 = .02). A significant positive linear age effect, however with small effect size, was found for μ (F(1,282) = 10.37, p = .001, ηp2 = .04), with more accurate Figure 4.2 The variability of time productions (ms) corrected for infrequent long response times across age in the distinct classes responses in older than younger children. µ; accuracy of time productions (ms) (95% CI) Figure 4.1 The accuracy of time productions (ms) corrected for infrequent long response times in the distinct classes Normal class ADHD -only class ADHD(+ASD) class ASD(+ADHD) class Age (years) Infrequent long response times (τ) No significant class effect nor significant class*age interaction effect was found for the τ (F(3,282) = 1.53, p = .21, ηp2 = .02 and F(3,282) = 0.20, p = .90, ηp2 = .00, respectively). A significant positive linear effect of age with a medium effect Normal-class ADHD-only class ADHD(+ASD)-class ASD(+ADHD)-class Classes 102 ; variability of time productions (ms) class (M = 925 ms) (p = .002 and p = .025, respectively), while the accuracy of size and a significant effect of age2 with a small effect size were found for these infrequent long response times (F(1,282) = 36.42, p < .001, ηp2= .12 and F(1,282) 103 Chapter 4 = 5.50, p = .02, ηp2= .02, respectively), see also Figure 4.3. Findings indicated problems, purely ADHD-behavior without any comorbidity, or both ASD and reduced infrequent long response times in older compared to younger children. ADHD-symptomatology, with one more prominent than the other. In contrast to our hypotheses, the pure ADHD-class did not deviate from the Normal class on any of the motor timing abilities (mu, sigma and tau). In fact, motor timing difficulties were found only in classes where both ADHD and ASD-symptoms Normalized; infrequent long response times (ms) Figure 4.3 Infrequent long response times (ms) across age in the distinct classes were present. The ADHD(+ASD) and ASD(+ADHD) classes showed a reduced motor timing accuracy (i.e. increased underestimation) compared to the Normal class. In addition, younger children in the ADHD(+ASD) class had a reduced variability in motor timing when compared to younger children in the Normal and ASD(+ADHD) classes, a pattern which was diminished across older children. The finding that the pure ADHD-class did not deviate from the Normal class on the motor timing abilities may seem to contrast previous studies that used the same motor timing paradigm and found a tendency to underestimate time and an elevated motor timing variability in ADHD (Rommelse et al., 2008; Normal class ADHD -only class ADHD(+ASD) class ASD(+ADHD) class Thissen et al. (under revision); van Meel et al., 2005). This contrast is likely due Age (years) to differences in groups across the studies; children who were DSM-defined as ‘ADHD’ in previous studies may actually have suffered from comorbid ASD- Finally, as a check on the interpretation of our findings, these analyses were symptoms as well. Additionally, our pure ADHD-class may have milder problems repeated without unaffected siblings of ASD, ADHD and ASD+ADHD affected than those typically included in case-control studies, suggesting that children children in the Normal class. This resulted in minor changes in outcomes, which with only the more severe behavioral symptoms show motor timing deficiencies. could not explain the absence of a difference between the ADHD-only class and In our data, both the ADHD(+ASD) class (with highest ADHD symptoms), and the the Normal class. Thus, the presence of unaffected siblings of affected children in ASD(+ADHD) class (with highest ASD symptoms), differed from the pure ADHD the Normal class does not change the conclusions. and Normal classes. This shows that current motor timing results cannot merely be explained by high ADHD severity with ASD playing no role. This, because DISCUSSION The present study examined whether reduced motor timing accuracy, increased motor timing variability, and infrequent long response times are specific for ADHD, or -in part- due to comorbidity with ASD. We compared motor timing difficulties across four homogeneous classes derived from a clinic and populationbased sample. These homogeneous classes presented either no behavioral 104 ADHD symptom severity in the pure ADHD-class and the ASD(+ADHD) class was highly similar, while the former class showed no timing deficits. Furthermore, the current findings parallel our previous study which indicated that homogeneous classes with children suffering from both types of symptoms were cognitively more impaired than children with pure ADHD symptoms, suggesting an overlapping cognitive background in ASD and ADHD (van der Meer et al., 2012). The current 105 Chapter 4 findings are well in line with studies that found deficits in time processing in profile of problems fits well with the symptom presentation of children with ASD children with ASD regardless of ADHD-comorbidity (Geurts et al., 2008; Maister & and comorbid ADHD, it follows that no claim can currently be made regarding the Plaisted-Grant, 2011). It has been suggested that deficits in temporal processing necessity of ADHD-symptoms for motor timing deficiencies to emerge when ASD- interact with primary symptoms such as the poor development of social cognition symptoms are present. in children with ASD (Falter & Noreika, 2011). Current underestimation of time across children and adolescents with both ASD and ADHD symptoms is commonalities in pure ADHD and ASD with comorbid ADHD can be further potentially also related to real-life difficulties in planning and organizing tasks and elucidated by analyzing brain-behavior relationships. The extent to which task completion. For example, children and adolescents with ASD and ADHD may substrates of motor timing related to pure ADHD are also related to ASD with perceive the time set for a given (school)task as very long, and may underestimate comorbid ADHD, and vice versa, can increase our understanding of the role of the time needed to complete tasks. time processing in the development of behavioral symptoms in ASD and ADHD. A recent meta-analysis on reaction time variability compared ADHD- It has been suggested that motor time processing deficits in ASD are due to affected children, adolescents (aged 13 to 18 years) and adults with clinical control an abnormal cortical connectivity and synchrony as well as more diffuse and groups (Kofler et al., 2013). Findings suggested that children but not adolescents widespread neural abnormalities, with reduced volumes reported in the parietal with ADHD had a slightly elevated variability compared to the clinical control lobe, limbic and cortical regions and white matter tracts (Belmonte et al., 2004; groups. In contrast, our findings suggest a reduced motor timing variability in the Gepner & Feron, 2009; Ivry, 2003).Functional magnetic resonance imaging (fMRI) class of youngest children with ADHD(+ASD) symptoms. This reduced variability data specifically focusing on the neural substrates of motor timing in children may reflect impulsivity or the tendency to rush a tedious task, one of the primary with ADHD indicated more confined deficits in the anterior cingulate gyrus, symptoms of ADHD. As also discussed by Falter et al. (in press), the interpretation supplementary motor area and their connections to fronto-striatal pathways of motor timing abnormalities in ASD and ADHD is obscured by the variety of tasks, (Rubia et al., 2009). Future fMRI studies across empirically defined homogeneous modalities, exposure durations and classifications used across studies. Our study ASD and ADHD classes may be better apt to inform us on not only the neural adds important knowledge to this topic by reducing the clinical heterogeneity mechanisms of motor timing, but also the possible shared and distinct neural present in DSM-defined ASD and ADHD group comparisons. Our comparisons of substrates of motor timing in pure ADHD and comorbid ASD and ADHD. motor timing abilities in empirically defined homogeneous ASD and ADHD classes suggest that ASD-symptoms contribute to motor timing abnormalities. However, in the three affected classes, whereas they were underrepresented in the Normal the role of ADHD in these combined classes is unclear, since a) no homogeneous class. This is because symptoms of ASD and ADHD are more frequently seen class with only ASD-symptoms emerged from the latent class analyses, and b) in boys than in girls (Kramer et al., 2008). Note that this overrepresentation was the class with most severe ADHD-symptoms presented with ASD-symptoms as present in all affected classes, and therefore did not affect comparisons between well (van der Meer et al., 2012). In addition, the classes that presented with ASD- those classes. Second, the latent classes were based on questionnaires. In symptoms also suffered from more symptoms on other behavioral domains such comparison with clinical interviews, questionnaires tend to overestimate the as oppositional behavior, fear, perfectionism and emotional lability. Although this degree of comorbidity, as questionnaires do not allow for further probing 106 Evaluation of the significance of motor timing differences and There are some limitations worthy of note. First, boys were overrepresented 107 Chapter 4 or explanation of questions (Tourangeau et al., 2000). However, a possible overestimation of comorbidity cannot account for the differences in motor timing abilities in the distinct latent classes (van der Meer et al., 2012). Third, the nature of our samples might have prevented us from detecting a homogeneous class with pure ASD-behavior. ASD without ADHD symptoms might be underrepresented in clinic-based samples and rare in population-based samples. Therefore, very large samples are required to examine this issue further. 108 109 Homogeneous combinations of ASD-ADHD traits and their cognitive and behavioral correlates in a population-based sample Jolanda M. J. van der Meer, Martijn G. A. Lappenschaar, Catharina A. Hartman, Corina U. Greven, Jan K. Buitelaar, Nanda N. J. Rommelse Journal of Attention Disorders (in press) 110 Abstract ASD and ADHD are assumed to be the extreme manifestations of continuous heterogeneous traits that frequently co-occur. This study aims to identify subgroups of children with distinct ASD-ADHD trait profiles in the general population, using measures sensitive across both trait continua, and aims to show how these subgroups differ in cognitive functioning. We examined 378 children (6-13 years) from a population-based sample. Latent class analyses (LCA) detected three concordant classes with low (10.1%), medium (54.2%) or high (13.2%) scores on both traits, and two discordant classes with more ADHD than ASD characteristics (ADHD>ASD, 18.3%) or vice versa (ASD>ADHD, 4.2%). Findings suggest that the ASD and ADHD traits usually are strongly related in the general population, and that a minority of children displays atypical discordant trait profiles characterized by differential visual-spatial functioning. This dissociation suggests that heterogeneity in ASD and ADHD is rooted in heterogeneity in the lower non-symptomatic end of the distribution. 113 Chapter 5 Autism Spectrum Disorder (ASD) and Attention-Deficit/Hyperactivity Disorder have been made in the identification of such discordant subtypes using LCA or (ADHD) are highly heritable developmental disorders that frequently co-occur LCGA (Latent Class Growth Analyses), adopting a longitudinal perspective (St. (Ames & White, 2011; Ronald et al., 2008). Twin studies reveal a moderate degree Pourcain et al., 2011; van der Meer et al., 2012; as described in chapter 3). St. of phenotypic overlap between ASD and ADHD both throughout the whole range Pourcain and colleagues (2011) suggested that children with ADHD symptoms of scores and at the upper extreme end (Reiersen et al., 2008; Ronald et al., without ASD symptoms may more often have a childhood-limited form of ADHD, 2008), and there is evidence for shared etiological factors for ASD and ADHD while children having both ASD and ADHD symptoms may have more persistent (Rommelse et al., 2010; 2011; St. Pourcain et al., 2011). Further, clinical studies ADHD symptoms, and possibly are more resistant towards treatment. Furthermore, have documented poor social skills, language delay, sensory overresponsivity, we previously used both ASD and ADHD clinical symptom measures as well attention problems, oppositional defiant behavior and emotion regulation problems as measures for comorbid internalizing and externalizing problems in LCA and in both ASD and ADHD (Gadow et al. 2009; Mulligan et al., 2009; Rommelse et identified amongst others two mutually exclusive classes of children with clinical al., 2011). An increasing number of studies showed an overlap between ASD and symptoms of both ASD and ADHD. One of these classes had proportionally more ADHD with respect to cognitive functions (Booth et al., 2010, Corbett et al., 2009, ASD than ADHD symptoms and the other class just the other way round (van Fine et al., 2008), and therefore, studying ASD and ADHD together may provide der Meer et al., 2012). Importantly, these classes showed opposite visual-spatial the most optimal strategy in examining both shared and unique underpinnings. processing capacities, suggesting the identification of behavioral subtypes may For an in-depth discussion on differing models of co-existence of ASD and ADHD, increase our understanding of the cognitive heterogeneity in both disorders as see also Banaschewski et al. (2007) and Rommelse et al. (2011). well as the etiology of their co-occurrence. ASD and ADHD are both highly heterogeneous disorders; however the A frequently overlooked issue that may have hindered progress in optimal approach to describe this heterogeneity remains unclear. Latent class identifying more homogeneous, etiologically distinct disorder subgroups is analyses (LCA) have been used with the aim to identify more homogeneous that not only ASD and ADHD populations, but also general populations are subgroups of both traits. This approach in the separate fields of ASD and ADHD characterized by heterogeneity. General populations are usually described with research previously resulted in the identification of subgroups that mainly differ a lack of precision and lumped together into a single group without symptoms. by disorder severity rather than in truly distinct categories in a general population This hinders the study of heterogeneity in this group, and ignores strong evidence sample (Acosta et al., 2008; Constantino, 2011; Volk et al., 2009). Recent studies that ASD and ADHD as well as other internalizing and externalizing behavioral using both ASD and ADHD symptom measures also disclosed concordant classes disorders exist on a continuum (Constantino, 2011; Levy et al., 1997; Lundstrom differing mainly in severity (Reiersen et al., 2007; Mulligan et al., 2009). Such et al., 2012; Plomin et al., 2009; Robinson, Munir, et al., 2011). Hence, cognitive concordant ASD-ADHD trait profiles highlight the shared etiology of both traits, and symptom heterogeneity at the upper end of the symptom distribution (i.e. with both disorders sharing a common genetic and biological basis. In contrast, in the clinical range) may well be reflective of similar cognitive and symptom discordant ASD-ADHD trait profiles, that are highly symptomatic on one trait but not heterogeneity at the lower end of the distribution. Such cognitive heterogeneity the other, may have atypical underpinnings. These underpinnings may translate was recently examined in a sample of ADHD-affected children and typically into differential prognoses and susceptibility towards treatment. Some successes developing children (Fair et al., 2012). Individual-based analyses on a range of 114 115 Chapter 5 cognitive tasks revealed that some of the cognitive heterogeneity in children with ADHD seemed to be nested within the variation in typically developing children: studying it using measures that are sensitive assessments across the continuous largely similar cognitive subtypes (i.e. neuropsychological subgroups) were ASD and ADHD traits. LCA were used to identify distinct ASD-ADHD profiles in revealed in both populations. The authors also showed that diagnostic accuracy the general population, and these profiles were examined for their internalizing increased somewhat when the ADHD versus control contrast was made within and externalizing problem and cognitive correlates. Given the correlations each cognitive subtype instead of whole group analyses. Furthermore, the study usually reported between both quantitative traits and the comorbidity between showed that a large part of the previously unexplained cognitive heterogeneity ASD and ADHD as extreme ends, it was expected that mostly concordant ASD- within ADHD seems actually not to be related to ADHD as a disorder (i.e. the ADHD classes would be detected that differed quantitatively but not qualitatively. upper end of the symptom distribution), but more so to cognitive heterogeneity However, by providing greater resolution of scores across the trait continua, we also present in the non-clinical part of the ADHD spectrum. also expected to have greater power to identify discordant classes with distinct behavioral and cognitive profiles. Ordinary assessments of psychopathology would not disclose such In the present study our focus is on the population-based sample, heterogeneity across the continuous ASD and ADHD traits, as they give resolution only to the affected end of disorders, reflected in skewed distributions of symptom METHODS measures. An exception are questionnaires that provide greater resolution across Participants the entire distribution, taking into account difficulties as well as possible strengths such as well-developed social-communication and attention traits. This approach seemed quite promising in distinguishing ADHD subtypes in population data on the Strengths and Weaknesses of ADHD symptoms and Normal behavior (SWAN) rating scale (Arcos-Burgos et al., 2010), which revealed latent classes with less than average hyperactivity and impulsivity. While the former study focused on symptom data, further work on identifying distinct non-affected subtypes may also provide us with a better understanding of the previously unexplained cognitive heterogeneity in ASD and ADHD. Therefore, the present study aimed to identify distinct ASD-ADHD trait profiles that are typified by distinct cognitive and behavioral profiles across the ASD and ADHD trait continua, including the non-symptomatic ends. These children were also part of our previous study in a combined clinical and population-based sample, where more than 80 % of them were lumped into ‘normal classes’ on the basis of clinical ADHD and ASD measures (van der Meer et al., 2012). The study has been approved by the Committee on Research involving Human Subjects (CMO) and children were enrolled between January 2009 and July 2011. Eligible children were 378 children from a random population cohort study (Schoolkids Project Interrelating DNA and Endophenotype Research; SPIDER). All children were between 6 and 13 years of age (M (SD) 8.9 (1.7), % male = 49.5). All were of Caucasian descent and had an estimated total IQ of at least 70 on the Wechsler Intelligence Scale (WISC-III) (Wechsler, 2002). Exclusion criteria were epilepsy, known genetic or chromosomal disorders, brain damage, and problems with vision or hearing. After complete description of the study to the parents, written informed consent from all parents was obtained. Measures ASD and ADHD symptom measures ASD and ADHD trait measures according to parents were obtained using the Autism Quotient (AQ) (Baron-Cohen et al., 2001; Hoekstra et al., 2008) which provides a quantitative measure of ASD-like traits in the general population, and 116 117 Chapter 5 the Strengths and Weaknesses of ADHD symptoms and Normal behavior (SWAN) rating scale (Hay et al., 2007), respectively. Both measures have shown adequate reliability and validity (Arnett et al., 2011; Hoekstra et al., 2008). Both are based on a Likert-type rating scale and show scores that followed a continuous distribution in the general population (Baron-Cohen et al., 2001; Polderman et al., 2007). The distribution of these measures resembled a poisson distribution rather than a normal distribution. Therefore, the subscales were modeled as count variables in the latent class model (Muthén & Muthén, 2006). Cognitive measures Table 5.1 Description of the cognitive measures Task Measurement potential Dependent variable(s) Baseline Speeda,b Speed and variability of motor output as response to external cue Mean reaction time (ms) and variability (SD of reaction time in ms) Facial Emotion Recognitiona,b Capacity to identify the facial emotional expression of happiness, sadness, anger and anxiety. Mean reaction time (ms) and accuracy on four emotions Response Organization Objectsa,b Motor Inhibition Difference in percentage of errors or mean reaction time (ms) between compatible and incompatible trials Cognitive Flexibility Difference in percentage of errors or mean reaction (ms) between compatible trials and mixed compatibleincompatible trials. Visuo-Spatial Attention Number of correct reproduced sequences in identical (forward) order Visuo-Spatial Working Memory Number of correct reproduced sequences in reversed (backward) order Verbal Attention Number of correct reproduced digits in identical (forward) order Verbal Working Memory Number of correct reproduced digits in reversed (backward) order Visual pattern recognition Number of correct and timely completed geometric designs The six neurocognitive tasks analyzed in this study have been described elsewhere (van der Meer et al., 2012; see chapter 3), a brief description of all cognitive dependent variables is provided in Table 5.1. Ceiling effects did not occur on any of the tasks as indicated by boxplot analyses on raw data (not Visuo-Spatial Sequencinga,b presented). Digit Spana,c Block Patternsa,c Note. a van der Meer et al.(2012) see also chapter 3; bde Sonneville (1999); c Wechsler (2002). Other internalizing and externalizing problems In addition to the normally distributed ASD and ADHD trait measures, two questionnaires measuring clinical symptom levels of ASD, ADHD, oppositional behavior, emotional lability, anxiety, perfectionism and psychosomatic complaints were obtained. These were the Social Communication Questionnaire (SCQ, Lifetime version; parent ratings) and the Conners’ Parent Rating Scale-Revised: Long version (CPRS-R:L), both validated instruments for screening developmental problems (Conners et al., 1998a, Rutter et al., 2003). 118 119 Chapter 5 Procedure The tasks described were part of the neuropsychological assessment battery used in the SPIDER project. Children completed the battery in approximately two hours and the order of task administration was counterbalanced. Due to time constraints, not all tasks were administered to all children. Full-Scale IQ was prorated by four subtests of the WISC-III; Block Design, Picture Completion, Similarities and Arithmetic. These subtests are known to correlate between .90 and .95 with Full-Scale IQ (Groth-Marnat, 1997; Kaufman, 1994). Parents were invited to fill in the aforementioned questionnaires concerning their child’s behavior. inherently confounded with symptoms of ASD and ADHD, and could therefore not be separated from the effect of class (Dennis et al., 2009). Class*age, class*sex, age*sex and class*age*sex interaction effects were examined and reported if significant. If non-significant, interactions were dropped from the model. Dependent variables were speed and/or accuracy measures for each task separately, or subscale scores on the internalizing and externalizing problems. All dependent variables were successfully normalized and standardized into z-scores by applying a Van der Waerden transformation (SPSS version 20). Some of the outcome measures were mirrored, so that the scores of all variables would imply the same: a higher z-score was indicative of a better performance. Correction for Data analyses To identify homogeneous ASD-ADHD trait classes, LCA were used on the subscale outcomes of the AQ, ranging from 0 to 30 (social skills, attention switching, local details, communication, imagination) and the subscale outcomes of the SWAN (inattention and hyperactive-impulsive). Subscale scores on the SWAN, ranging from 9 to 63, were mirrored so that the scores on all subscales would imply the same: a higher score was indicative of more symptoms. The LCA were carried out using Mplus version 6.11 (Muthén & Muthén, 2006). Both the probability for a multiple comparisons was applied according to the False Discovery Rate (FDR) controlling procedure with a p-value setting of .05 (Benjamini, 1995). Effect sizes were defined in terms of percentage of variance explained (ηp2). Small, medium and large effects were defined in variances of .01, .06 and .14 respectively (Cohen, 1988). RESULTS Identifying Homogeneous Symptom Classes child to belong to each of the classes and the conditional probabilities for children The LCA on the AQ and SWAN subscales were based on fit and accuracy in a particular class to show specific behavior were estimated. Next, children were measures (Nylund et al., 2007), and revealed a solution with five classes. Five admitted to the class with the highest probability. Mean subscale sum scores on classes had the best fitting BIC and SSA BIC values, and entropy (.887), and a the seven aforementioned subscales were computed, and presented in a line bootstrapped lo-mendell-rubin likelihood ratio test p-value < .001 (see also Table chart, so that quantitative differences between classes could be examined. Size 5.2). This, combined with the most informative class profiles and all correlation and significance of class differences on these subscales were determined with a matrix probabilities > .900, indicated accurate classification. The AQ and SWAN MANOVA. profiles of the classes are presented in Figure 5.1. For the purpose of simplicity, Next, class differences with respect to age and sex were analyzed to the classes were labeled. Three concordant classes emerged which had either check for possible confounders. The identified classes were examined regarding low, medium or high levels of both ASD and ADHD traits (see also Table 5.3). their cognitive profiles and their internalizing and externalizing problems We refer to those as ‘LL’ (Low ASD, Low ADHD; 10.1%), ‘MM’ (Medium ASD, separately, using ANCOVA’s with class-membership as a fixed factor, and Medium ADHD; 54.2%) and ‘HH’ (High ASD, High ADHD; 13.2%), and two age and sex as covariates. IQ was not implemented as a covariate since IQ is discordant classes with either higher scores on the ADHD traits than on the ASD 120 121 An Figure 5.1 Class scores on AQ (left) and SWAN (right) subscales trait ‘ASD>ADHD’ (4.2%). The LL-class scored low on both the AQ and the SWAN, 0,5 the MM-class scored intermediate on both measures and the HH-class scored 0,4 -0,1 SWAN. Roughly 30% of the children in the ASD>ADHD class passed the clinical -0,3 cut-off for the ASD-measure. Again all scores were below clinical cut-off on the -0,4 Class General tests of model fit Entropy .878 BIC 16558.88 SSA BIC 16511.29 Technical output VLMR LRT p-value LMR adj. LRT p-value .00 .00 -Im pu At te n ls iv e tio n tio n in a ag H yp er ac tiv e Im C om Lo m ca un ic a lD et ai ls Sk ills At te n Table 5.2 Results of latent class analyses on AQ and SWAN measures tio n -0,5 ia l have been undisclosed in the ordinary assessments of ASD and ADHD. Class 4: HL (4.2%) n = 15 So c ASD and ADHD clinical symptom scales, indicating that these distinctions would Class 2: LM (18.3%) n = 68 -0,2 in g while the ASD>ADHD-class scored relatively high on the AQ and low on the Class 5: HH (13.2%) n = 46 0 Sw itc h The ADHD>ASD-class scored intermediate on the SWAN and low on the AQ, 0,1 tio n scored below the clinical cut-off on the ASD and ADHD clinical symptom scales. Class 3: MM (54.2%) n = 203 0,2 Mean sum scores the HH-class passed the clinical cut-off for both measures. Still, all three classes Class 1: LL (10.1%) n = 38 0,3 relatively high on the AQ as well as the SWAN. Roughly 30% of the children in 2 Pe r behavioral domain trait ‘ADHD>ASD’(18.3%), or higher scores on the ASD trait than on the ADHD No. fe Ps ctio yc n co ho m so pl ai nt s nt er ac om m un ic a St er e ot So y ci al pr ob l In at te R es tle ss n C pr ogn ob H l em yp er ac t O pp os Em iti ot io na ll a li C So ci a Chapter 5 Class 1: Normal 41.6 % cognive domain Note. Social skills, attention switching, local details, communication and imagination are subscales of the AQ (Autism Spectrum Quotient), attention and hyperactive-impulsive are subscales of the SWAN (The Strengths and Weaknesses of ADHD symptoms and Normal behavior scale). A higher mean factor sum score indicated that children in that class lacked population-based sample more competences or showed more problems on the specific domain. 1,5 1,0 .857 16179.36 16106.39 .24 .00 4 .862 16017.76 15919.40 .03 .00 speed(21.9 %) as dependent variables revealed that, as expected, the five classes differed 0 5 .887 15912.57 15788.83 .34 .00 medium accuracy-high overall-0,5significantly (p< .001). Next, all classes were pairwise compared on the 6 .862 15839.38 15690.26 .09 .00 separate -1,0 ASD and ADHD subscales. Only 11 out of 70 comparisons did not reach a mean sum score 3 A MANOVA using class as a fixed factor and the ASD and ADHD subscales 0,5 high accuracy-medium speed (24.2 %) low accuracy-medium speed (35.3 %) significance. Roughly, the non-significant differences were on ASD-measures -1,5 on on gn iti co re re re co co gn iti gn iti on y or m ng ng em or la tte em m te at nt io n y io n nt pu t ro ut ve rb al ot ot o or o ut pu t low accuracy-low speed (on the left side of Figure 5.1) between either the LL and ADHD>ASD class or (18.6 %) n n ot ot io io rn em em su sp at ia lw pa tte or ki ia os pa t ki or w al ve rb of y lit va ria bi ee d of m m between the ASD>ADHD and HH class, or on ADHD-measures (on the right mean of d sp ee ra cy of su al vi cu vi su o- vi side of Figure 5.1) between the LL and ASD>ADHD class. The distribution of all sp Note. Entropy refers to classification accuracy, BIC refers to Bayesian Information Criterion, SSA BIC refers to Sample Size Adjusted BIC, VLMR LRT refers to the vuong-lo-mendell-rubin likelihood ratio test, LMR adj. LRT refers to the bootstrapped lo-mendell-rubin likelihood ratio test. a From a 6 classes solution onwards, the p-value may not be trustworthy due to local maxima. ac children across the distinct classes, as well as the sex, age, and IQ distributions cognive domain are provided in Table 5.3. Boys were overrepresented in the classes with higher levels of ASD and/or ADHD traits (classes HH and ASD>ADHD), whereas girls clinic-based sample were overrepresented in the class with the lowest levels of both traits (class LL). 1,5 122 ean sum score 1,0 0,5 0 123 high accuracy-medium speed (16.5 %) medium accuracy-high Chapter 5 When corrected for the influence of age and sex, no changes in differences between the classes were found. Figure 5.2 Differences between the classes on measures of block patterns and visual-spatial working memory Block patterns 1,0 0,8 To test in which cognitive domains the classes overlapped or differed, separate 0,6 ANCOVAs were used for each cognitive domain, with age and sex as covariates. 0,4 in their block pattern performance (F (4,376) = 5.61, p< .001, ηp2 = .06). The ASD>ADHD class showed superior block pattern performance compared to the ADHD>ASD class, while the ADHD>ASD class also performed worse compared -0,2 -0,4 to the MM-class. The HH-class performed comparable to the ADHD>ASD class, and worse than the ASD>ADHD class. Next, the overall class effect reached -0,8 trend-level significance for visual-spatial working memory (F (4,348) = 2.04, p -1,0 = .09, ηp2 = .02), significant post-hoc results did not survive the correction for Other Internalizing and Externalizing Problems of the Distinct Classes concordant classes LL (n=38) MM (n=204) discordant classes HH (n=50) ADHD>ASD (n=69) ASD>ADHD (n=16) 0,8 0,6 0,4 0,2 Z-score more accurate responses classes did not differ on the other cognitive measures studied. p = .001 p = .017 Visual-spatial working memory and the ADHD>ASD class indicated a significant difference in visual-spatial performance and visual-spatial working memory are presented in Figure 5.2. The p = .026 1,0 multiple comparisons. Still, pairwise comparison between the ASD>ADHD class discordant classes were relatively small. Overall class effects on block pattern p = .005 0,0 -0,6 working memory (F (1,79) = 5.94, p = .02, ηp2 = .07), despite the fact that both p = .003 p = .001 0,2 Z-score The discordant classes differed from the concordant classes and from each other more accurate responses Cognitive Profiles of the Distinct Classes 0,0 -0,2 -0,4 -0,6 Scores on the clinical questionnaire (CPRS) indicated that all classes represent the non-symptomatic side of the continuum: none of the classes scored in the clinical range on any of the subscales (i.e. oppositional behavior, emotional lability, anxiety, perfectionism and psychosomatic complaints). The concordant classes with intermediate or relatively high scores on the ASD and ADHD traits also presented elevated scores on the other internalizing and externalizing traits. For the discordant classes, the highest levels of the other internalizing and externalizing problems were present in the ASD>ADHD class. In contrast, the 124 -0,8 -1,0 concordant classes LL (n=32) MM (n=188) HH (n=49) discordant classes ADHD>ASD (n=66) ASD>ADHD (n=14) Note. The means are adjusted for the covariate age. Group differences presented were based on a mean age of 8.9 years. Error bars represent 1 standard error. LL refers to the concordant class with low levels of ASD and ADHD symptoms, MM refers to the concordant class with intermediate scores on both traits, HH refers to the concordant class with high levels of both symptoms. ADHD>ASD refers to the discordant class with intermediate levels of the ADHD and low levels of ASD symptoms, ASD>ADHD refers to the discordant class with high levels of ASD symptoms and low levels of ADHD symptoms. The overall class effect reached trend-level significance for visual-spatial working memory, (p = .09, ηp2 = .02), and significant post-hoc differences in visual-spatial working memory did not survive FDR-correction. 125 Chapter 5 Table 5.3 Demographic characteristics of the children in the distinct classes Concordant classes Discordant classes LLa MMa HHa ADHD> ASDa ASD> ADHDa n=38 (10.1%) n=205 (54.2%) n=50 (13.2%) n=69 (18.3%) n=16 (4.2%) Contrasts based on p-values of .05 Switching, Local Details, Communication, Imagination), the clinical cut-off of the total score on the AQ in children is 76 (Auyeung et al., 2008). d The total score on the SCQ (Social Communication Questionnaire) reflected the total amount of ASD-symptoms. e The mirrored total scores on the SWAN (The Strengths and Weaknesses of ADHD symptoms and Normal behavior scale) reflected the degree of ADHD-related symptoms. f Subscale scores on the CPRS (Conners’ Parent Rating Scale) subscales reflected the degree of domain-specific symptoms. The official cut-off for clinically relevant symptoms is a subscale score above 63. M (SD) M (SD) M (SD) M (SD) M (SD) Age in years 9.3 (1.6) 8.8 (1.8) 9.1 (1.7) 8.5 (1.7) 9.7 (1.5) ns ADHD>ASD class did not present increased scores on the other internalizing % Male 28.9 47.8 76.0 40.6 75.0 LL = ADHD>ASD = MM < ASD>ADHD = HH and externalizing traits. When corrected for the influence of age, sex, class*age, 101.9 (9.5) 104.1 (8.5) 110.4 (12.0) HH <ASD>ADHD class*sex and class*age*sex interaction effects, results did not change. Results Estimated full 105.7 (9.3) 106.1 (10.2) scale IQb are also presented in Table 5.3. ASD measures Total score AQc 29.7 (8.1) 43.0 (7.9) 74.4 (15.7) 21.9 (5.8) 70.1 (11.3) LL <ADHD>ASD< MM < ASD>ADHD = HH Total score SCQd 2.8 (2.8) 4.3 (3.2) 9.1 (5.7) 2.2 (2.2) 7.9 (4.3) LL = ADHD>ASD < MM < ASD>ADHD = HH 41.8 (7.6) 68.1 (9.7) 82.8 (12.2) 68.1 (8.4) 44.3 (8.2) LL = ASD>ADHD < MM = ADHD>ASD < HH population-based sample. As hypothesized, the individual-based analyses T-score CPRSe 44.4 (3.9) ADHD 50.4 (8.9) 62.1 (10.9) 47.6 (8.2) 45.6 (5.5) all< HH revealed mostly quantitatively differing, concordant ASD-ADHD classes (77.5%) ADHD measures Total score SWANe with either low, medium or high scores on both traits, and two discordant Oppositionalf 46.1 (6.0) 49.6 (7.7) 59.8 (11.1) 46.8 (7.1) 48.1 (5.3) all< HH Emotional Labilityf 43.6 (4.4) 47.3 (7.7) 55.9 (10.9) 44.2 (6.1) 47.3 (8.4) LL = ADHD>ASD<MM = ASD>ADHD< HH Anxietye 46.0 (6.0) 50.0 (9.2) 59.3 (11.6) 45.2 (5.8) 53.7 (12.0) LL = ADHD>ASD < MM = ASD>ADHD< HH Perfectionism 45.5 (4.9) 47.5 (7.2) 54.8 (11.6) 43.2 (4.4) 56.1 (10.2) LL =ADHD>ASD< MM <ASD>ADHD= HH 47.2 (5.5) 49.8 (9.7) 58.1 (14.1) 47.8 (8.0) 48.8 (7.3) Psycho somatic Complaintse all< HH Note. a LL is the class with low levels of ASD and ADHD symptoms, MM refers to the class with intermediate levels of ASD and ADHD, and HH is the class with relatively high levels of ASD and ADHD symptoms. ADHD>ASD is the class with intermediate levels of ADHD symptoms and low levels of ASD symptoms, ASD>ADHD refers to the class with relatively high levels of ASD symptoms and low levels of ADHD symptoms. b Full-scale IQ was estimated by four subtests of the WISC-III (Wechsler, 2002): Block Design, Picture Completion, Similarities and Arithmetic. These subtests are known to correlate .90 to .95 with Full-scale IQ (Groth-Marnat, 1997). c The total score on the AQ (Autism Spectrum Quotient) reflected the total amount of ASD-symptoms (subscales Social Skills, Attention 126 The present study examined whether differentiated ASD-ADHD latent classes typified by distinct cognitive and behavioral profiles can be identified in a Other internalizing and externalizing problems e DISCUSSION classes with either more ADHD symptoms than ASD symptoms (18.3%), or more ASD symptoms than ADHD symptoms (4.2%). When comparing the latter two discordant classes, the specific combination of ASD>ADHD was characterized by a superior visual-spatial processing, whereas the ADHD>ASD combination was characterized by inferior visual-spatial processing. Furthermore, the class with elevated scores on both traits (HH) presented a cognitive profile which closely resembled the profile of the ADHD>ASD class. The elevated levels of internalizing and externalizing problems in the classes with either high scores on both traits or more ASD symptoms than ADHD symptoms did not translate into more performance deficits than in the other classes. Intriguingly, many of the findings in these non-clinical ASD-ADHD classes were an extension of our previous study on clinical ASD-ADHD classes as well as other studies across clinical ASD and ADHD samples (Mulligan et al., 2009; 127 Chapter 5 Reiersen et al., 2007; St. Pourcain et al., 2011; Todd et al., 2002; van der Meer et as recently discussed by Fair and colleagues (2012), our findings suggest that al., 2012). A first parallel was that one discordant ADHD>ASD profile was much heterogeneity in clinical developmental disorders are rooted in comparable more common (18.3%) than the other discordant profile ASD>ADHD (4.3%); the heterogeneity present in the general population. Given that previously more than latter profile even remained undisclosed in our previous study. This may suggest 80% of our population sample was lumped together into ‘normal’ classes (van that across the general population, just as across the clinical population, most of der Meer et al., 2012), the current findings can be seen as an extension of those the children who express ASD-behavior also present the less severe ‘precursor’ of findings across the non-symptomatic end of the ASD and ADHD traits. ADHD-behavior, as has been hypothesized previously (Rommelse et al., 2011). A second resembling finding was that the ASD>ADHD class is again characterized dissociation in visual-spatial functioning (block pattern performance and working by superior visual spatial functioning, whereas the ADHD>ASD class is memory) between classes displaying either more ASD traits than ADHD traits or characterized by inferior visual spatial functioning. Thus, superior visual-spatial vice versa, which was also found previously for clinical cases. This dissociation functioning in children with higher scores on the ASD trait than the ADHD trait might pinpoint towards differential organization and/or functioning of neural holds across the general population and the clinical population alike. This finding substrates underlying visual-spatial information processes that are oppositely is in keeping with recent studies across the general population reporting that a involved in ASD and ADHD pathology. While normative visual-spatial attention higher level of autistic traits measured with the AQ across the general population is biased towards global processing (i.e. global interference), children with was also associated with an enhanced visual working memory (Grinter et al., ASD are said to have a visual perceptual processing style that facilitates local 2009; Richmond et al., 2012). A third similarity was that the class with high scores rather than global processing (Booth et al., 2003; Happé & Frith, 2006). Such on both traits (class HH) showed a cognitive profile that most closely resembled local processing is favorable in the completion of cognitive tasks like the block the ADHD>ASD class. This may suggest that children with elevated scores on pattern task and embedded figures test (EFT). Individuals with ASD or high levels both traits primarily suffer from cognitive problems with an ADHD-alike etiology. of mild ASD-like traits often show a superior performance in such tasks since Alternatively, the behavior in the HH-class may actually reflect a true co-occurring no global perceptual bias needs to be surpassed (Grinter et al., 2009; Shah & of cognitive features of both ASD and ADHD, but the prime cognitive deficits Frith, 1983). A reduced global-to-local perceptual process in ASD has also been specific for ASD may be obscured by the cognitive deficits most robustly found in approved in recent fMRI studies (Just, Keller, Malave, Kana & Varma, 2012; Liu, ADHD. A fourth parallel was that boys were overrepresented in the classes with Cherkassky, Minshew & Just, 2011; McGrath et al., 2012), which may indicate higher levels of ASD and/or ADHD, which corresponds with the upper extreme less top-down control and increased local connectivity in ASD. Such increased ASD and ADHD traits being more easily recognized in boys than in girls (Kramer local connections were not found in studies in ADHD; neural activity patterns et al., 2008). This may indicate that sex-differences at the upper extreme end of rather indicated a frontal, striatal and parietal hypofunction in ADHD (Bush, Valera phenotypic traits are also embedded in more typically developing children, as & Seidman, 2005; Silk et al., 2005; Vance et a., 2007). The widespread reduced has been discussed previously (Neuman et al., 2005). This finding may suggest top-down control described in those studies is not specific for visual processes that sex differences in clinical referral and diagnoses of ASD and ADHD are not in ADHD, and may pinpoint towards an overall reduced attentional network. We based on a clinical bias, but rather reflects a true predisposition in males. Hence, aim to follow-up on these findings by comparing children with elevated scores on 128 Of particular relevance is the extension into the general population of the 129 Chapter 5 either ASD or ADHD, children with elevated scores on both, and control children regarding brain activation patterns during visual-spatial task performances. This study was not without limitations. First, information on all phenotypic domains relied on parent reports. Compared to clinical interviews, surveys tend to overestimate the degree of co-occurrence of ASD and ADHD, since the degree of response variability that can be measured is limited (Tourangeau et al., 2000). This limitation may have affected the distribution of children across the classes in favor of the concordant classes and may have hampered the disclosure of discordant classes. Therefore, the optimal approach for studying these traits is the use of structured psychiatric interviews. Second, questionnaires measuring the symptom levels of the other internalizing and externalizing traits are not designed to examine the lower extreme end of the phenotypic spectrum. This however, did not impede differentiation across the classes, as can be seen in differences among the classes in other internalizing and externalizing problems (see also Table 5.3). Third, apart from the visual-spatial functioning, the classes did not differ on the other cognitive domains (see also Table 5.1). This may be due to the weakened associations between the cognitive measures on the one hand and the reduced range of scores on both ASD and ADHD traits on the other. Fourth, as aforementioned, sex differences were profound across all classes, with boys overrepresented in the classes with higher levels of the ASD (and ADHD) traits and girls overrepresented in the class with the lowest levels of both traits. However, we do not believe this has affected the results, since the effect of sex was analyzed and, when necessary, accounted for in the study. In sum, the present study showed that, in the general population, 77.5% of the children presents with concordant ASD-ADHD trait profiles, while 22.5% of the children displays atypical discordant trait profiles characterized by differential visual-spatial functioning. This dissociation was previously also reported in classes with clinical symptoms of ASD and ADHD, suggesting that heterogeneity in ASD and ADHD is rooted in heterogeneity present in the non-symptomatic end of the distribution. 130 131 Using cognitive profiles to examine the relationship between ASD and ADHD Jolanda M. J. van der Meer, Catharina A. Hartman, Jan K. Buitelaar*, Nanda N. J. Rommelse* *shared last author Under review 132 Abstract The objective was to examine if segmenting in cognitively homogeneous classes is a useful approach in detecting shared and unique mechanisms underlying Autism Spectrum Disorder (ASD) and Attention-Deficit/Hyperactivity Disorder (ADHD). Therefore, latent class analyses (LCA) were performed on motor speed and variability, verbal and visual-spatial attention, verbal and visual-spatial working memory, visual pattern recognition and emotion recognition in 360 children from a population-based sample and 254 children from a clinic-based sample (5 - 17 years). Classes were compared on several behavioral symptom scales. LCA in the population and clinic-based samples revealed a similar four class solution typified by qualitatively different speed-accuracy trade-offs: high accuracy-medium speed (21.9% of the population sample and 16.5% of the clinic sample), medium accuracy-high speed (24.2% and 24.4%), low accuracy-medium speed (35.3% and 39.0%) and low accuracy-low speed (18.6% and 20.0%). These classes were respectively associated with lowest and highest levels of ASD and ADHD symptoms in the clinical sample, with an overall strong predictive effect. Associations with clinical symptoms were much weaker in the population sample. Classes were not characterized by domain specific cognitive strengths or weaknesses. Cognitive subtyping appears a powerful strategy to uncover the mechanisms underlying ASD and ADHD. The relevance of cross-domain generic cognitive factors fits current models of abnormal neural connectivity in ASD and ADHD. The weak associations between cognition and behavior in the population sample suggest that cognitive functioning may only predict behavior when other risk or protective factors are present. 135 Chapter 6 Autism Spectrum Disorder (ASD) and Attention-Deficit/Hyperactivity Disorder anxiety and motor problems. Preliminary evidence further suggested exposure to (ADHD) are neuropsychiatric developmental disorders that frequently co-occur perinatal factors may differ between these ASD-ADHD classes. Finally, St. Pourcain (for review, see Rommelse et al., 2010). The frequent comorbidity of both and colleagues (2011) used Latent Growth Curve Analyses (LGCA) to examine disorders is likely due to a substantial overlap in genetic factors and functional to what degree changes in ASD and ADHD symptoms during development are and structural brain characteristics between ASD and ADHD (for review, see associated. They showed that children with ADHD symptoms persisting into late Rommelse et al., 2011). Disclosing these shared underpinnings is complicated adolescence were more ASD symptomatic compared to children with other ADHD because both disorders encompass multiple distinct subtypes with overlapping symptom trajectories. Overall, these studies indicate that empirically defined symptom presentations. A basic assumption is that specific genetic deficits, brain ASD-ADHD classes show partially distinct patterns of comorbid pathology and abnormalities or cognitive impairments may underlie these disorders only in a distinct developmental trajectories, and suggest they may be differentially linked subgroup of the patients (Brieber et al., 2007; Maher, 2008; Veatch, Veenstra- to etiological factors. Vanderweele, Potter, Pericak-Vance & Haines, 2014; Verté et al., 2006). Hence, an important strategy is to empirically segment this heterogeneous group of on ASD, ADHD and comorbid symptom data (van der Meer et al., 2012, as individuals with ASD, ADHD or a combination of ASD and ADHD into subgroups described in chapter 3). Our results pointed to at least two ASD-ADHD comorbid with possibly a more homogeneous set of underlying mechanisms. subtypes that could be differentiated by the pattern of ASD and ADHD symptoms. The merits of empirically defining more homogeneous disease subtypes Subsequently, we were able to document that each class was also characterized have already been demonstrated in separate studies of ASD and ADHD. For by a quite distinct cognitive profile. That is, one ASD-ADHD class showed poor instance, subtypes with a homogeneous symptom profile defined by Latent recognition of facial emotions in combination with superior visual spatial (working Class Analyses (LCA) show less heterogeneity with respect to age (Elia et al., memory) skills, whereas the other ASD-ADHD class showed normal abilities 2009), comorbid symptoms (Acosta et al., 2008; Beuker et al., 2013), associated in recognizing facial emotions, yet poor visual spatial (working memory) skills. cognitive deficits (Fair et al., 2012; Munson et al., 2008), and prognosis (St. These findings suggest cognitive functions may be used in our search for more Pourcain et al., 2011). These subtypes are also possibly more stable across homogeneous classes of ASD-ADHD patients. informants (Althoff et al., 2006) than DSM based subtypes. Three previous studies used LCA or an equivalent technique to study empirically defined ASD- symptoms is potentially even more promising in identifying distinct ASD-ADHD ADHD classes. Reiersen and colleagues (2007) described several LCA defined classes. Cognitive performances can be measured more objectively than ADHD classes with distinctive degrees of increased levels of ASD symptoms. clinical symptoms, and is potentially more closely linked to the neurobiological Particularly the most severely impaired ADHD class showed the highest levels underpinnings (Gottesman & Gould, 2003). A proof of concept study on of ASD symptoms. Mulligan and colleagues (2009) extended these results by neuropsychological heterogeneity in typically developing children and children showing that children with ADHD combined type can be subdivided into several with ADHD reported that children with ADHD could be segmented in four classes with different levels of ASD symptoms. These ASD-ADHD classes also subgroups characterized by specific cognitive difficulties for response variability, showed distinct patterns of comorbid pathology, such as oppositional behavior, executive functioning, temporal information processing and signal detection 136 We recently took this line of research a step further by performing LCA Segmenting based on cognitive performance rather than on clinical 137 Chapter 6 (Fair et al., 2012). As these subgroups did not differ in ADHD symptoms, the III) or Wechsler Adult Intelligence Scale (WAIS-III) (Wechsler, 1989; 2000; 2002). clinical phenotype of ADHD may be rooted in multiple distinct cognitive subtypes. Exclusion criteria were epilepsy, known genetic or chromosomal disorders, brain Moreover, the cognitive subgroups disclosed across typically developing children damage, and problems with vision or hearing. The studies have been approved largely resembled the cognitive subgroups disclosed in ADHD. This suggests by the local Committee on Research involving Human Subjects (CMO). After the that part of the heterogeneity in ADHD is nested in normal variation. The broader study procedures had been fully explained, informed consent was signed by all implication is that cognitive profiles disclosed through a bottom-up approach are participants, whereby parents signed informed consent for children younger than generic, i.e. not only relevant for normal development and ADHD but possibly 12 years of age. also for other neurodevelopmental disorders such as ASD. Therefore, the present study aimed to extend these lines of research Measures by examining if homogeneous cognitive segmenting is a useful approach in A large variety of cognitive domains was assessed, robustly associated with ASD detecting shared and unique cognitive substrates for ASD and ADHD. Based on (i.e. emotion recognition and visual pattern recognition) or ADHD (i.e. motor our previous results (van der Meer et al., 2012), we hypothesized that a cognitive speed and variability, verbal and visual attention, and verbal and visual–spatial subtype might be identified with superior visual spatial skills and inferior emotion working memory), as documented in previous studies (Booth & Happé, 2010; recognition abilities that was most strongly linked to ASD; and a cognitive subtype Corbett et al., 2009; Fine et al., 2008; Rommelse et al., 2011; van der Meer et with inferior visual-spatial skills and normal emotion recognition abilities that was al., 2012). These cognitive domains are also summarized in Table 6.1. Children most strongly linked to ADHD. completed the assessment battery in approximately two hours and the order of METHODS task administration was counterbalanced. Due to time constraints, not all tasks were administered to all children. Full-Scale IQ was prorated by four subtests of Participants the WPPSI, WISC-III or WAIS-III; Block Design, Picture Completion, Similarities Between January 2009 and July 2011, 360 eligible children were recruited from and either Vocabulary (BOA) or Arithmetic (SPIDER) (Wechsler, 1989; 2000; a random population cohort study (Schoolkids Project Interrelating DNA and 2002). These subtests are known to correlate between .90 and .95 with Full-Scale Endophenotype Research; SPIDER) and 254 children from a clinical ASD-ADHD IQ (Groth-Marnat, 1997). Parents were invited to fill in several questionnaires genetic study (Biological Origins of Autism; BOA). The SPIDER cohort consisted concerning their youngster’s behavior. of the full range of children who attend primary schools, the BOA cohort consisted of ASD, ADHD and ASD+ADHD affected children and their non-affected siblings (for a full description, see Box 1.2 regarding the study samples, or see van Steijn et al., 2012). Both samples aimed to cover wide behavioral and cognitive trait distributions. All children were between 5 and 17 years of age, of Caucasian descent and had an estimated total IQ of at least 70 on the Wechsler Preschool and Primary Scale of Intelligence (WPPSI), Wechsler Intelligence Scale (WISC- 138 139 Chapter 6 Table 6.1 Description of the cognitive measures population-based sample, behavioral symptom measures (parent reports) were also taken from the Strengths and Weaknesses of ADHD symptoms and Normal Task Measurement potential Dependent variable(s) Baseline Speeda,b Speed and variability of motor output as response to external cue Mean reaction time (ms) and variability (SD of reaction time in ms) Digit Spana,c Verbal Attention Number of correct reproduced digits in identical (forward) order Data analyses Verbal Working Memory Number of correct reproduced digits in reversed (backward) order To identify homogeneous cognitive classes, latent class analyses (Mplus version Visuo-Spatial Attention Number of correct reproduced sequences in identical (forward) order Visuo-Spatial Working Memory Number of correct reproduced sequences in reversed (backward) order Block Patternsa,c Visual pattern recognition Number of correct and timely completed geometric designs Facial Emotion Recognitiona,b Capacity to identify the facial emotional expression of happiness, sadness, anger and anxiety. Mean reaction time (ms) and accuracy on four emotions Visuo-Spatial Sequencinga,b Note. a van der Meer et al. (2012) see also chapter 3; b de Sonneville (1999); c Wechsler (1989; 2000; 2002). Procedure Behavioral symptom measures (parent reports) were based on the Social Communication Questionnaire (SCQ, Lifetime version), the Conners’ Parent Rating Scale-Revised: Long version (CPRS-R:L) and the Autism Quotient (AQ) in both samples. The SCQ and the CPRS-R:L are both validated instruments for screening ASD and ADHD (Conners et al., 1998a, Rutter et al., 2003) while the AQ provides quantitative measures of ASD-like traits in the general population (Baron-Cohen et al., 2001; Hoekstra et al., 2008). For children in the clinic-based sample scoring above clinical cut-offs on the SCQ and CPRS (DSM-subscales), the Parental Account for Childhood Symptoms (PACS) and Autism Diagnostic Interview Revised (ADI-R) were administered by a certified clinician to obtain a diagnosis of ADHD and/or ASD (Le Couteur et al., 2003; Taylor et al., 1991). In the 140 behavior (SWAN) rating scale that provided a quantitative measure of possible strengths in the general population (Hay et al., 2007). 6.11, Muthén & Muthén, 2006) were conducted in both samples separately on the following 9 cognitive measures: speed and variability of motor output, verbal attention, verbal working memory, visuo-spatial attention, visuo-spatial working memory, visual pattern recognition and speed and accuracy of identification of facial emotions. The cognitive measures were corrected for the influence of age by calculating age regressed residuals, and next successfully normalized and standardized into z scores by applying a Van der Waerden transformation. Cognitive measures were uncorrected for sex and IQ as both are inherently confounded with and therefore cannot be separated from ASD and ADHD (Dennis et al., 2009). The measures for speed (response time) and variability were mirrored, so that for all variables a higher score was indicative of a better performance. The mean factor sum scores of all cognitive domains were computed and presented in line charts, so that quantitative differences between classes could be examined. Size and significance of differences were determined with multivariate analyses of variance (MANOVAs), after which the correction for multiple comparisons was applied according to the false discovery rate (FDR) controlling procedure with a p-value setting of .05 (Benjamini, 1995). Secondly, we examined whether the cognitive subtypes differed for ASD and/or ADHD symptoms, comorbid oppositional behavior, cognitive problems, anxiety, perfectionism, psychosomatic complaints and emotional lability symptoms. To this end, we ran MANOVAs with the cognitive classes as independent variable, and all behavioral domains as dependent variables in both samples separately. Polynomial contrasts tested for linear and quadratic 141 Chapter 6 class differences regarding symptom data. Effect sizes were defined in terms of off (medium accuracy-high speed). Class 3 (35.3% and 39.0%, respectively) was Cohen’s d, mall, medium and large effects were defined as explained variances of best described as low accuracy-medium speed. Class 4 could be viewed as low .01, .06 and .14 respectively (Cohen, 1988).Correction for multiple comparisons accuracy-low speed (18.6% and 20.0%, respectively). The characteristics of the was again applied according to the FDR-controlling procedure with a p-value classes are provided in Table 6.2. Supplemental separate analyses of affected and setting of .05 (Benjamini, 1995). Only the effects that remained significant after unaffected individuals from the clinic-based sample confirmed an LCA solution of the FDR correction were reported. four classes typified by the same qualitatively different speed-accuracy trade-offs (see Supplement 6.1). In a supplementary analysis we examined the representativeness and stability of the cognitive profiles. Therefore children from the clinic-based (BOA) sample were divided into two categories: affected if the ADI-R and/or PACS scores were above clinical cut-offs, and unaffected if, despite familial risk, these scores were below clinical cut-offs. Latent class analyses were conducted on the same cognitive measures (separately normalized and standardized) in both subsamples, as in our main analyses. RESULTS Identifying Homogeneous Cognitive Classes in Both Samples Latent class analyses were based on fit and accuracy measures and visual inspection of the figures (Nylund et al., 2007). This revealed a solution with four classes in both samples. Four classes had the best fitting BIC values and satisfying entropy (.719 for the population-based sample and .775 for the clinic-based sample), combined with the most informative class profiles. The cognitive profiles of the classes are presented in Figures 6.1a and 6.1b. For ease of interpretation, the classes were labeled according to their cognitive profiles. Given the emerged similarities in the cognitive profiles across the clinic and population-based samples, these labels were applicable to classes from both samples. These classes were typified by qualitatively different speed-accuracy trade-offs rather than strengths or weaknesses on specific cognitive domains. Class 1 (21.9% of the population-based sample and 16.5% of the clinic-based sample) was best referred to as high accuracy-medium speed, whereas class 2 (24.2% and 24.4%, respectively) showed an opposite speed-accuracy trade- 142 143 of sp e -1,0 -1,5 ed of m v m va ot a ot or ria or ria oubi ou bi lit lit t tp p y uyt o ut of fm m ot ot or or ou ou tp tp ut v ut ve er rb ba al la ve atve tte rb terb nt al nt al io w io w no n or rk ki in ng g m m vi evmisu em su ooro-s or spvi vi y p y su atsu at oiaoia l asp la sp tteat tte at ia ntial nt lw io w io n or n or ki ki ngvi ng vi su su m m al al em em pa a ac oprat or ttecc cu y te y rnura rn ra recy re cy co o co of gnf e gn em iti m iti onot on ot sp iosp io ne n ee e r r e e d cdo o co of gnf e gn em iti m iti onot on ot io io n n re re co co gn gn iti iti on on ed sp e mean sum mean score sum score of mean sum mean score sum score 1,5 0,5 a. 1,0 b. 144 1,0 0 high accuracy-medium speed(21.9 %) 0,5 -0,5 0 -1,0 medium accuracy-high high accuracy-medium speed (24.2 %) speed(21.9 %) -0,5 -1,5 low accuracy-medium medium accuracy-high speed (35.3 %) speed (24.2 %) -1,0 -1,5 1,5 0,5 1,0 0 0,5 -0,5 0 -1,0 -0,5 -1,5 low accuracy-low speed low accuracy-medium (18.6 %) speed (35.3 %) mean low accuracy-low speed (18.6 %) cognive domain mean cognive domain 1,5 clinic-based sample clinic-based sample high accuracy-medium speed (16.5 %) medium accuracy-high high accuracy-medium speed (24.4 %) speed (16.5 %) low accuracy-medium medium accuracy-high speed (39.0 %) speed (24.4 %) low accuracy-low speed low accuracy-medium (20.1 %) speed (39.0 %) mean low accuracy-low speed (20.1 %) cognive domain mean Note. A higher mean sum score indicated that children in that class had more competencies or showed less problems on the specific cognitive domain. The cognitive measures were cognive domain corrected for the influence of age. a 49.2 T-score CPRS Social problems 47.8 49.5 30.7 31.5 T-score CPRS Inattention T-score CPRS Hyperactivity SWAN Attention SWAN Hyperactivity/ Impulsivity ADHD measures 4.5 43.6 Total score AQ 8.8 8.0 8.1 8.5 7.9 18.8 4.0 34.0 32.7 51.7 49.0 48.9 43.2 4.5 8.5 108.6 111.2 8.0 50.6 1.6 49.4 9.5 M SD M Total score SCQ ASD measures Estimated IQ % Male Age in years Class 2 n= 87 Class 1 n = 79 53.5 9.2 M n=127 Class 3 3.8 7.1 7.0 10.2 8.8 8.0 20.3 4.2 33.4 34.8 53.6 52.3 50.3 43.7 38.8 8.6 M 7.8 8.1 10.7 10.4 8.8 18.3 3.7 32.8 33.4 52.4 51.2 50.9 39.4 4.5 8.5 103.1 1.7 SD n=67 Class 4 9.3 M 7.4 7.4 9.8 9.7 9.2 15.8 3.8 n.s. 1<3 1<3 1<3 n.s. n.s. n.s. 54.2 53.1 57.7 58.1 8.7 11.5 11.5 12.4 29.6 8.9 58.8 56.6 58.7 60.1 9.7 10.9 102.6 2.0 SD Class 2 n = 62 63.6 12.4 M 13.3 13.3 14.3 32.0 9.0 63.3 60.4 65.0 69.1 11.4 10.0 103.1 1.9 SD n = 99 Class 3 66.7 11.4 M 14.3 12.4 17.0 32.7 9.4 68.6 64.8 69.4 78.1 15.7 12.0 100.4 3.0 SD n = 51 Class 4 All < 1 n.s. 2 < all 12.9 12.1 16.4 n/a n/a 1=2< 4& 1<3 1 = 2 < 4& 1<3 1=2<4 31.1 1 = 2 < 4 10.8 1 = 2 < 4 12.4 4.3 SD Class Class low accuracy- low accuracymedium low speed speed Clinic-based sample Class medium accuracyhigh speed 58.1 11.6 M n = 42 Class 1 Class high accuracymedium speed 71.4 n.s. 2=4 <1 Contrasts based on p-values of .05 3=4 9.9 116.1 < 2 = 1 2.1 SD Class Class low accuracy- low accuracymedium low speed speed 8.1 101.7 1.2 SD Class medium accuracyhigh speed Population-based sample Contrasts based on p-values of .05 population-based sample Class high accuracymedium speed Figures 6.1a and population-based 6.1b Latentsample cognitive classes across population and clinic1,5 based samples 1,0 hi C om m un ic iv eIm pu ls i At te n H ti yp Im ag in at i at i Lo ca lD et a Sw itc So ci al Sk i At te nt io n H yp er ac t cognive domain Table 6.2 Demographic characteristics and behavioral symptoms in the distinct cognitive classes m sp ee d of m otva ot or ri or ria oaubi ou bi lit lpit t tp y y u ut of t of m m ot ot or or ou ou tp tp ut v ut ve e rb rb al al ve v at e at terb te rb nt al nt al io w io w no n or r ki ki ng ng mv m vi emisu em su ooro-s or spvi vi y p y su atsu at oiaoia l asp la sp tteat tte at ia ntial nt lw io w io no n or rk ki i ngv ng vi i su su m m al al em em pa a ac oprat or ttecc y te y cu rnura rn ra recy re cy co o co of gnf e gn em iti m iti onot on ot sp iosp io ne n ee reed re d co o co of gnf e gn em iti m iti onot on ot io io n n re re co co gn gn iti iti on on va d ee sp Chapter 6 cognive domain 145 146 Note. a Full Scale IQ was estimated by four subtests of the Wechsler Preschool and Primary Scale of Intelligence (WPPSI), Wechsler Intelligence Scale (WISC-III), or Wechsler Adult Intelligence Scale (WAIS-III) (Wechsler; 1989; 2000; 2002). To prevent a restriction of variance in symptom measures in the population-based sample, normally distributed ASD (AQ) and ADHD (SWAN) trait measures were included. SWAN-scores were mirrored so that the scores on all questionnaires would imply the same: a higher score was indicative of more symptoms / less favorable outcome. 1<4 13.7 59.6 56.1 55.2 9.9 8.9 8.6 47.3 T-score CPRS Emotional lability 8.0 48.1 8.2 47.2 47.2 n.s. 50.1 12.7 14.1 1<4 62.6 57.6 57.3 12.3 9.6 11.7 48.8 T-score CPRS Psychosomatic complaints 9.0 49.8 8.8 51.4 50.0 n.s. 53.2 14.7 15.0 16.0 n.s. 11.9 55.3 53.9 52.9 9.6 7.6 7.6 48.1 T-score CPRS Perfectionism 7.8 48.0 9.9 47.8 47.1 n.s. 51.3 11.5 12.8 n.s. 16.4 63.4 60.5 57.2 16.5 8.4 11.5 49.8 T-score CPRS Anxiety 8.8 48.6 7.4 51.4 49.2 n.s. 56.1 14.3 14.8 1=2< 4& 1<3 11.6 62.7 59.2 55.2 9.8 9.5 10.4 48.7 T-score CPRS Cognitive problems 8.1 49.1 8.6 53.2 51.7 1=2 <3 52.8 11.8 10.8 1< 3 =4 62.0 59.2 56.3 9.3 9.0 49.7 T-score CPRS Oppositional behavior Comorbid measures 8.5 50.2 8.8 50.2 8.6 49.0 n.s. 51.7 13.3 13.8 12.8 Chapter 6 MANOVAs using class as fixed factor and all cognitive domains as dependent variables revealed that the four classes differed significantly on all cognitive measures in the clinic-based sample as well as the population-based sample (both p’s < .001). In FDR-corrected post hoc comparisons for the clinic-based sample, 11 of 54 comparisons did not reach significance, for the populationbased sample, 16 of 54 comparisons did not reach significance. Roughly, the nonsignificant differences were found on comparisons with the low accuracymedium speed class in both the clinic and population-based sample, and in the population-based sample on comparisons between all classes regarding the accuracy of identification of facial emotions. Behavioral Profiles of the Homogeneous Cognitive Classes Next, we examined whether the cognitive classes differed in ASD, ADHD and comorbid symptoms, by running MANOVAs for all behavioral domains in both samples separately. The results are also presented in Table 6.2. We found linear class contrasts for severity of ASD, ADHD, oppositional behavior, cognitive problems (all p-values < .001), psychosomatic complaints (p = .01) and emotional lability (p = .001) in the clinic-based sample. The high accuracy-medium speed class had the lowest amount of symptoms, the low accuracy-low speed class the highest amount of symptoms (d’s between .58 and .69 for ASD-measures, between .91 and .96 for ADHD-measures and between .62 and .86 for comorbid problems). In contrast, effect sizes were much smaller in the population-based sample, with linear class contrast only present for ADHD symptoms and cognitive problems (both p-values < .005), with only a small difference in symptom severity between the high-accuracy-medium speed class and low accuracy-low speed class (average d = .28). Results are presented in Figures 6.2a and 6.2b. 147 Chapter 6 Figure 6.2 Differences between the cognitive classes on a) measures of ASD and ADHD, and b) oppositional behavior, cognitive problems, anxiety, perfectionism, psychosomatic complaints and emotional lability symptoms Clinic-based sample ,50 ,50 ,30 SCQ AQ CPRS Hyperactivity CPRS Inattention ,10 -,10 -,30 Z-score (effect size) ,70 more behavioral problems Z-score (effect size) more behavioral problems Clinic-based sample ,70 -,50 -,70 ,70 ,30 ,10 -,10 -,30 -,50 high accuracy medium accuracy medium speed high speed low accuracy medium speed -,70 low accuracy low speed low accuracy low speed ,10 SCQ AQ CPRS Hyperactivity CPRS Inattention SWAN -,10 -,30 -,50 Z-score (effect size) ,50 more behavioral problems Z-score (effect size) more behavioral problems medium accuracy low accuracy high speed medium speed ,70 ,30 -,70 high accuracy medium speed Population-based sample Population-based sample ,50 a. Oppositional Cognitive problems Anxiety Perfectionism Psychosomatic Emotional lability ,30 ,10 Oppositional Cognitive problems Anxiety -,10 Perfectionism Psychosomatic Emotional lability -,30 -,50 high accuracy medium accuracy low accuracy medium speed high speed medium speed low accuracy low speed b. -,70 high accuracy medium accuracy low accuracy medium speed high speed medium speed low accuracy low speed Note. Error bars represent 1 standard error. SCQ = Total score on the Social Communication Questionnaire, AQ = Total score on the Autism Quotient. ADHD and comorbid scores were based on the Conners’ Parent Rating Scale-Revised (CPRS-R:L), SWAN = Total score on the Strengths and Weaknesses of ADHD symptoms and Normal behavior rating scale, population sample only. 148 149 Chapter 6 DISCUSSION This study examined if segmenting ASD and ADHD into homogeneous cognitive classes is a useful approach in detecting shared and unique substrates for ASD and ADHD. Our main finding is that LCA in a population and a clinic-based sample revealed similar four class solutions typified by qualitatively different speed-accuracy trade-offs: high accuracy-medium speed, medium accuracy-high speed, low accuracy-medium speed and low accuracy-low speed. These classes were respectively associated with lowest and highest levels of ASD and ADHD (and several comorbid) symptoms in the clinical sample, with an overall strong predictive effect. Effects were much weaker or absent in the population sample. Of note, classes were not characterized by domain specific cognitive strengths or weaknesses. In the clinic-based sample the speed-accuracy trade-off pattern was strongly linked to between-class differences in ASD and ADHD symptom severity. Children with inaccurate and slow performance across a range of tasks (i.e. blue line in Figure 6.1) had the highest levels of ASD and ADHD as well as comorbid symptoms, while children performing accurate at a normal pace showed the lowest levels of ASD, ADHD as well as comorbid symptoms. The comparison of these two most extreme groups resulted in moderate to large class differences. This is in sharp contrast to previously reported effect sizes (small to moderate at best) when comparing measures of cognition in DSM-defined groups (for reviews see Gargaro, Rinehart, Bradshaw, Tonge & Sheppard 2011; Taurines et al., 2012). This seems to indicate that cognitive subtyping results in more homogeneous groups than DSM-based categories, and may therefore be a powerful strategy in examining causal factors underlying ASD and ADHD. The profile associated with the lowest amount of ASD and ADHD (and comorbid) symptoms (i.e. the ‘protective’ profile) was also associated with higher intelligence and somewhat older age. This may reflect a trade-off favoring accuracy over speed that is associated with brain maturational processes. In contrast, the profile associated with the highest amount of ASD, ADHD and 150 comorbid symptoms (i.e. the ‘risk’ profile) was characterized by slow and inaccurate performances across a range of tasks. Clearly, children in this latter group had difficulty with any cognitive task that was presented to them (regardless of the specific aim of measurement), even tasks with very low cognitive demands (i.e. simple baseline speed). These ‘generally at risk’ and ‘generally favorable’ cognitive profiles may suggest that generic cognitive skills strongly influence cognitive performances across domains in a consistent manner. It is tempting to speculate about the neural basis of these generic neurocognitive factors in ASD and ADHD by linking them to abnormalities in brain connectivity (Di Martino et al., 2013). Compromised small-world properties, i.e. reductions in local efficiency and modularity within several functional networks, have previously been reported for both ASD and ADHD (Rudie et al., 2013; Wang et al., 2009; Wass, 2011), although findings are still somewhat inconclusive, and may be cognitive statedependent (for reviews, see de la Fuente, Xia, Branch & Li, 2013; Vissers, Cohen & Geurts, 2012). Future studies might examine how the current generic cognitive profiles map into the efficiency and modularity of functional neural networks. In any case, our findings indicate that a task-transcending cognitive impairment, resulting in an overall slower and more inaccurate performance, is an aspecific but strong predictor of psychiatric dysfunctioning. This calls into question the use of extensive cognitive batteries that are now often used for research and clinical purposes, when performance across these tasks is mainly driven by a generic cognitive impairment (Kuntsi et al., 2010; Millan et al., 2012) that may also be measurable using much shorter paradigms. The absence of cognitive profiles with domain specific strengths/ weaknesses contrasts with the idea of multiple cognitively different developmental pathways to ADHD as suggested in several theory papers (for review, see Nigg et al., 2005). At first sight, findings may also seem in contrast to those of our previous study that described two classes of children with ASD and ADHD with some distinctiveness in their cognitive profiles regarding emotion recognition and visual spatial functioning (van der Meer et al., 2012). However, this may be due 151 Chapter 6 to the reversal of our dependent and independent variables in both approaches. In the previous study we started to derive classes with homogeneous symptom samples, one highly enriched in ASD/ADHD affected individuals and one profiles and then detected virtually no cognitive differences between both groups, population derived sample. One the one hand, the derived cognitive subtypes with the exception of the two aforementioned domains. In contrast, in the current were very similar in both samples, indicating that the LCA solution was fairly study we first constructed cognitively homogeneous classes, and then identified robust and not driven by sampling bias. On the other hand, the predictive value of large between-class differences in clinical symptoms. The large degree of shared cognitive subtype regarding symptom measures was much stronger in the clinic variance was apparently more powerful and relevant for the formation of cognitive sample compared to the population sample. It remains puzzling how this can subtypes than the cognitive parameters (i.e. in visual spatial functioning and be explained. A restriction of variance in symptom measures in the population emotion recognition abilities) that were differentially related to ASD and ADHD sample is unlikely to account for this, since we used amongst others also normally (van der Meer et al., 2012). distributed ASD and ADHD traits. However, when comparing both samples for Our findings do contrast sharply though with those reported by Fair et each cognitive subtype separately (see Supplement 6.2), it can be noticed that al. (2012). Their findings suggested that subgroups of ADHD-affected children the ‘risk’ profile (low accuracy and low speed) in the clinic sample performs have specific strengths or weaknesses in distinct cognitive domains involved in overall somewhat worse than the ‘risk’ profile in the population sample, whereas the disorder. It remains to be seen why these outcomes diverge. Since substantial the ‘protective’ profile in the clinic sample (high accuracy and medium speed) overlap was present in the task batteries used in both studies (motor speed and performs overall somewhat better than the ‘protective’ profile in the population variability, verbal and visual spatial attention, and verbal and visual spatial working sample. In other words, there may be less cognitive variance in the population memory), it seems unlikely that the content of the current cognitive battery explains sample compared to the clinical sample, which may explain the (much) lower the absence of cognitive subtypes with domain specific strengths/ weaknesses. predictive value regarding behavioral measures, as is also found in another recent A prominent difference between the two studies that is more likely to explain study (Rajendran, Rindskopf, et al., 2013). This may suggest cognitive functioning the difference in results is the use of either community detection (CD) analyses only predicts behavior when other risk or protective factors are present and that (Newman, 2006) or latent class analyses (LCA), both are techniques that aim to in every day life, favoring speed over accuracy or the other way around is overall derive homogeneous subgroups. Current data were also examined with the use equally adaptive. of CD. This however resulted in very unstable outcomes in split-half analyses. This instability in outcomes led us to prefer the LCA approach, providing replicable was younger than the clinic-based sample, and age has proven to be an important results in both samples as well as within the clinic sample when split into an factor in the pattern of association between ASD and ADHD (St. Pourcain et al., affected and non-affected subsample. In any case, our findings challenge the 2011). Therefore, the cognitive measures were corrected for the influence of currently wide held view of multiple different (sometimes viewed as independent; age by calculating age regressed residuals. Furthermore, the cognitive profiles de Zeeuw, Weusten, van Dijk, van Belle & Durston, 2012; Sonuga-Barke, 2005) were replicated in the clinic and population-based sample despite these age- developmental pathways to ADHD, by indicating that these pathways are not differences, which underscores that the current patterns of association between independent, but likely to be related to underlying generic cognitive impairment. ASD and ADHD could not be explained by age-differences. Second, we had to 152 We took the opportunity to perform our analyses in two independent This study was not without limitations. First, the population-based sample 153 Chapter 6 be selective in choosing the cognitive measures used for constructing cognitive Supplemental Material homogeneous classes, and different cognitive tests and paradigms might have Identifying Homogeneous Cognitive Classes Within the Clinic-Based been selected. In particular, extension of our findings by including other ASD and/ Sample or ADHD related measures such as Theory of Mind abilities, and processing of social and non-social reward (Ames & White, 2011) may be worthwhile. In conclusion, cognitive subtypes -defined by different speed-accuracy trade-offs instead of domain specific strengths and weaknesses- are similar in clinic and population-based samples and strongly related to ASD, ADHD and several other symptom domains in the clinic sample. Cognitive subtypes did not (or much weaker) relate to behavior in the population-based sample, suggesting cognitive functioning may only predict behavior when other risk or protective factors are present. In two additional latent class analyses it was examined if the four cognitive profiles found in the clinic-based sample could also be identified in cases and unaffected siblings, respectively. These subsamples consisted of either affected children (n = 144) or their siblings who had ASD and ADHD symptoms below clinical ADI-R and PACS cut-offs (n = 110). The cognitive profiles of the four classes solutions in these subsamples are presented in Supplement 6.1a and 6.1b. These profiles closely resembled the cognitive profiles in the complete clinic-based sample, indicating that the cognitive profiles used are stable and representative for the complete ASD and ADHD continuum from no problems to mild and up to severe problems. 154 155 Chapter 6 Supplement 6.1a and 6.1b Latent cognitive classes across clinic-based subsamples Supplement 6.2 Class comparisons between population and clinic-based samples mean sum score Clinic-based affected subsample High accuracy – medium speed 1,5 1,0 0,5 high accuracymedium speed(15.3%) 0 -0,5 medium accuracyhigh speed(22.2 %) -1,0 sp ee d va of ria m bi ot lit or y of ou m tp ot ut or ve ve ou r rb ba tp al ut la w tte or n v k vi tio is su uo ing n om s sp em pa at tia or ia y la lw vi ac t te or su cu nt k a in io ra lp g n cy at m te of e m r n or sp of e re y co m ee o g d t n i o of iti on em n re co ot gn io n iti re on co gn iti on -1,5 low accuracy-medium speed (40 %) low accuracy-low speed (22.2 %) mean cognitive domain a. 1,5 1,0 0,5 0 -0,5 -1,0 -1,5 * * Medium accuracy – high speed 1,5 1,0 0,5 0 -0,5 -1,0 -1,5 Low accuracy-medium speed Clinic-based unaffected subsample mean sum score 1,5 1,0 0,5 high accuracy-medium speed (25.5 %) 0 -0,5 medium accuracy-high speed (38.2 %) -1,0 sp e va ed of ria m bi ot lit or y of ou m tp ot ut or ve ve ou rb rb tp al al ut w at or te k n in vi tio vi g su su n m ooe sp sp m at or at ia ia la y lw ac vi tte s o cu ua r nt ra io l p king cy n at m of te e r m of n sp or re em y ee co ot d g io ni of n t io em re n co ot gn io n i tio re n co gn iti on -1,5 b. 1,5 1,0 0,5 0 -0,5 -1,0 -1,5 * * low accuracy-medium speed (22.7 %) low accuracy-low speed (13.6 %) mean Low accuracy – low speed 1,5 1,0 0,5 0 -0,5 -1,0 -1,5 * cognitive domain Note. Affected subsample (a): ADI-R and PACS scores above clinical cut-offs. Unaffected subsample (b): ADI-R and PACS scores were below clinical cut-offs, despite familial risk of ASD (and ADHD). A higher mean sum score indicated that children in that class had more competencies or showed less problems on the specific cognitive domain. The cognitive measures were corrected for the influence of age. 156 157 Chapter 6 population sample clinic sample 1,0 0,5 att en tio n ve rba me l wor mo kin ry g al vis uo att -spa en tio tial n vis wo uo rki -sp ng a me tial mo ry vis ua rec l pa og tte nit rn ion ac cu e rac rec moti y of og on nit ion sp ee do rec f og emo nit ion tion -1,5 ve rb sp -1,0 va ria b -0,5 ility ou of m tpu t otor 0 ee d ou of m tpu oto t r Z-score more cognitive competencies 1,5 cognitive domain Note. A higher standardized score indicated that children in that class had more competencies or showed less problems on the specific cognitive domain. The scores were standardized for both samples together, cognitive measures were corrected for the influence of age. Significant differences were marked with an asterisk. 158 159 A randomized, double-blind comparison of atomoxetine and placebo on response inhibition and interference control in children and adolescents with autism spectrum disorder and comorbid attention-deficit / hyperactivity disorder symptoms Jolanda M. J. van der Meer, Myriam Harfterkamp, Gigi van de Loo-Neus, Monika Althaus, Saskia W. de Ruiter, A. Rogier T. Donders, Leo M. J. de Sonneville, Jan K. Buitelaar, Pieter J. Hoekstra, Nanda N. J. Rommelse Journal of Clinical Psychopharmacology, 2013; 33 (6), 824-827. 160 Abstract The aim of this study was to investigate whether atomoxetine led to improvement of response inhibition and interference control in children with Autism Spectrum Disorder (ASD) and Attention-Deficit/Hyperactivity Disorder (ADHD), and whether ADHD symptom improvement was mediated by improvement in response inhibition and interference control. Therefore, 97 children (6-17 years) with ASD and ADHD and a Total IQ of ≥ 60 were randomly assigned to double blind treatment with either atomoxetine or placebo for 8 weeks, followed by a 20 weeks openlabel phase. Neuropsychological assessments and behavioral questionnaires were completed at baseline, after the double blind phase and after the open-label phase. Atomoxetine treatment was associated with improvement of response inhibition but not of interference control. Next, atomoxetine decreased ADHD symptoms as assessed by parents and teachers, yet no significant correlation was found between improvement in measure of inhibitory control or ADHD symptoms. Atomoxetine improved response inhibition but not interference control in children with both ASD and ADHD, independent from progress in ADHD-symptomatology. These findings suggest that cognitive improvements and improvements on the behavioral level do not necessarily co-occur, and that distinct pathogenic pathways may play a role in the occurrence of clinical symptoms and cognitive dysfunctions. 163 Chapter 7 With prevalence rates of 1% for Autism Spectrum Disorder (ASD) and 5% for except for a pilot study (Arnold et al., 2006), well-controlled clinical trials of Attention-Deficit/ Hyperactivity Disorder (ADHD), these disorders are among the atomoxetine in ASD patients are lacking. Open-label studies suggest decreased most commonly diagnosed psychiatric syndromes in children (Baird et al., 2006; hyperactivity and improved attention and learning after atomoxetine treatment Polanczyk et al., 2007). Impairments in ASD are characterized by decreased (Charnsil, 2011; Fernández-Jaén et al., 2012; Hazell, 2007; Jou et al., 2005; Posey communication and social interaction skills, and repetitive and restricted behavior et al., 2006; Troost et al., 2006; Zeiner, Gjevik & Weidle, 2011). These effects of and interests. Impairments in children with ADHD reflect severe inattention, open-label studies need confirmation by double-blind, placebo-controlled studies hyperactivity and impulsivity impulsivity (American Psychiatric Association, 2013). to establish treatment benefits of atomoxetine on the ADHD-symptoms in ASD. Although the clinical descriptions of both disorders are quite distinctive, ASD and ADHD co-occur frequently: about 20 to 50% of the patients with ADHD meet criteria severity in children with both ASD and ADHD using a double-blind placebo- for ASD and about 30 to 80% of the patients with ASD meet criteria for ADHD controlled design in 97 children (Harfterkamp et al., 2012). Results indicated that (Reiersen & Todd, 2008; Rommelse et al., 2010). Despite the fact that the DSM- atomoxetine improved ADHD symptoms in children with ASD according to both IV prohibits a diagnosis of ADHD in the context of ASD, it has now widely been parent and clinician-based ratings and was generally well tolerated. Furthermore, recognized that both disorders can co-occur and share a substantial proportion effects were stronger for symptoms of hyperactivity-impulsivity compared to of etiological risk factors (Rommelse et al., 2010; 2011; Ronald et al., 2008; St. symptoms of inattentiveness, which is in accordance with the only placebo- Pourcain et al., 2011). Therefore, the DSM-5 draft includes the possibility of a controlled crossover pilot trial available (Arnold et al., 2006). All in all, previous comorbid diagnosis. This has given impulse to several clinical trials investigating open-label studies and our double-blind placebo controlled study suggest the effectiveness of pharmacological treatments such as methylphenidate for atomoxetine is indeed effective in reducing ADHD symptoms in patients with ASD ADHD symptoms in ASD-patients (Murray, 2010; Posey et al., 2007; Research and ADHD. Units On Pediatric Psychopharmacology, 2005; Santosh et al., 2006; Stigler, et al., 2004). Overall, it was concluded that the benefits seen from methylphenidate associated with ADHD provides more insight into the working mechanisms of are smaller amongst patients suffering from both ASD and ADHD-symptoms the pharmacological agent. For the selective norepinephrine re-uptake inhibitor as opposed to pure ADHD patients. Further, adverse effects such as agitation, atomoxetine, little is known about the cognitive working mechanisms. Several reduced appetite, gastrointestinal symptoms and fatigue were more frequently (animal-)studies suggest that inhibitory control, one of the core deficits associated reported amongst comorbid patients. These side effects are reportedly severe with ADHD symptoms (Barkley, 1997; Pennington & Ozonoff, 1996), improves enough to lead to discontinuation and suspended treatments. Therefore, recent with atomoxetine (Bari, Eagle, Mar, Robinson & Robbins 2009; Chamberlain et reviews on the efficacy and tolerability of medical treatment options for ADHD-like al., 2007; 2012; Faraone, Biederman et al., 2005; Gau & Shang, 2010; Robinson symptoms in individuals with ASD suggest that in addition to methylphenidate, et al., 2008), but others have found no effects (de Jong et al., 2009; Nandam et atomoxetine may be a reasonable choice to target ADHD-symptoms in ASD al., 2011) or even deleterious effects (Graf et al., 2011). In contrast, for the mixed (Benvenuto et al., 2012; Cortese, et al., 2012; Doyle & McDougle, 2012; for review dopamine and norepinephrine re-uptake inhibitor methylphenidate, research has see Ghanizadeh, 2012; Handen et al., 2011; Hanwella et al., 2011). However, consistently shown its improvement of the neural mechanisms of inhibitory control 164 We previously reported the effect of atomoxetine on ADHD symptom Knowledge of the effects of medical treatment on cognitive deficits 165 Chapter 7 in ADHD-patients (Ashare et al., 2010; Chamberlain et al., 2011; Groom et al., 2010; 2003). Moreover, all children were screened for the presence of ADHD-combined Lee, Han, Lee & Choi, 2010; Nandam et al., 2011; Rubia et al., 2011; Scheres et subtype (ADHD-C) and had to meet DSM-IV-TR criteria A through D for ADHD al., 2003). The contrasting findings for atomoxetine may partially be explained by (American Psychiatric Association, 2000). Both parents and children over the age the broad definition of inhibitory control. Inhibitory control encompasses amongst of 12 years had to give written informed consent after procedures and possible others inhibition of an ongoing response and interference control (Scheres et side effects were explained to them. Exclusion criteria included an IQ below 60 on al., 2004). Inhibition of an ongoing response (or response inhibition) is best a Wechsler Intelligence Scale (WISC-III)(Wechsler, 2002) a weight of less than 20 described as the ability to suppress pre-potent behavior that is inappropriate kg, presence of psychosis, bipolar disorder, substance abuse, a serious medical or no longer required, an output-process reflecting stimulus selection, whereas illness, history of seizures, ongoing use of psychoactive medications other than interference control refers to the cognitive control needed to prevent interference the study drug, and intended start of a structured psychotherapy or in-patient due to competition of relevant and irrelevant stimuli, an input-dependent process, treatment. Females who were post-menarche and sexually active had to take a reflecting response organization. These aspects do not necessarily correlate with pregnancy test to exclude pregnancy. The study was approved by the national each other, hence a selective norepinephrine re-uptake inhibitor (atomoxetine) and local institutional review board committees, and children were enrolled and a mixed dopamine and norepinephrine re-uptake inhibitor (methylphenidate) between October, 2006 and March, 2008. may influence distinct mechanisms of inhibitory control. This is in accordance with earlier findings that suggested both overlapping and distinct cognitive effects which 5 were excluded based on their intelligence quotient and/or cut-off scores of methylphenidate and atomoxetine (Chamberlain et al., 2011; Robinson et al., on the ADI-R. In total, 97 children (83 boys, 14 girls) were randomly assigned to 2008). It should be noted that these studies have focused on ADHD-only, healthy either placebo (n=49) or atomoxetine (n=48) treatment. Five children randomized controls or rodents. Change in the neural mechanisms of inhibitory control as a to atomoxetine and three children randomized to placebo discontinued during result of atomoxetine treatment has thus far not been studied in patients suffering the double blind treatment, see also Figure 7.1. Of the 89 children who completed from both ASD and ADHD. Our aims were to examine 1) whether atomoxetine the double-blind placebo controlled phase one child decided to stop and not to improved two forms of inhibitory control (response inhibition and interference continue in the open label treatment period. 88 started the open-label phase; control), and 2) whether ADHD symptom improvement (Harfterkamp et al., 2012) 42 previously on atomoxetine and 46 previously on placebo. Fifteen children was mediated by improvements in inhibitory control. withdrew from this phase of the study because of adverse events or lack of At the start of the study, 102 children were assessed for eligibility, after efficacy, resulting in 73 children completing the open-label phase. No differences METHODS between the placebo and atomoxetine group were found regarding age, sex or Participants IQ, see also Table 7.1. Eligible children were referred to one of nine participating Dutch medical centers across the country, between 6 and 17 years of age at their initial visit, and suffering from ASD and ADHD. ASD diagnoses were established by clinical assessment and at least two ADI-R subscale scores had to be above the clinical cut-off (Rutter, 166 167 Chapter 7 Table 7.1 Baseline demographic characteristics of subjects randomized to the placebo and atomoxetine group ATXa (n = 48) M (SD) Placebo group (n = 49) M (SD) F, p Age in years 10.5 (2.7) 10.4 (2.9) 0.01 NS Male (%) 87.5 83.7 0.16 NS WISC-III IQb 91.0 (16.4) 94.6 (17.7) 0.46 NS Note. ATX= Atomoxetine. In 2 children, estimated total intelligence quotient (IQ) was not available. 1 child (randomized to atomoxetine) had a nonverbal IQ above 60, but was nottestable with regard to his verbal IQ, and 1 child (randomized to placebo) had widely differing nonverbal and verbal IQs (of 55 and 85,respectively); therefore a total IQ could not be validly determined. a b Measures Both tasks contained an instruction trial wherein the examiner provided a typical item of the task, and a separate practice session (de Sonneville, 1999) Test–retest reliability and validity of the computerized ANT-tasks are satisfactory and have been described and illustrated elsewhere (de Sonneville, 2005). Response Inhibition The Go-No Go task was administered to measure the response inhibition of prepotent responses. In this task, 24 Go signals, in response to which the subjects had to press a key, were randomly mixed with 24 No Go signals, to which responses had to be suppressed. The Go signals were open squares, whereas Study Design The study consisted of three periods, starting with a screening/washout period with a duration of 3 to 28 days, followed by an 8 week double blind treatment with either atomoxetine or placebo in a 1:1 ratio and then a 20 week open label phase. During the first phase (T0), all children were screened for eligibility, including a diagnostic and medical evaluation. The first neuropsychological assessment was administered in order to correct for potential differences in previous test experience. At the start of the double blind phase (T1), children were randomly assigned to a double blind treatment with either atomoxetine or placebo capsules, the No Go signals were closed squares. In each trial the signal was preceded by a warning tone lasting 500 ms, signals were presented every 3000 ms. The valid response window was 200 ms to 2300 ms post stimulus onset. The primary outcome measure was the proportion of false alarms after No Go signals. The secondary outcome measures not specifically related to response inhibition were the proportion of missed Go signals, and the response time and response time variability of correct responses. Interference Control titrated to a fixed once-daily dose (first week: 0.5 mg/kg/day; second week: 0.8 The Focused Attention task was administered to measure interference control, mg/kg/day; then 1.2 mg/kg/day for six weeks). Atomoxetine and placebo were i.e., the degree of distractibility by irrelevant information. In this task, 20 relevant identical in appearance. The neuropsychological measures were administered target trials, in response to which the subject had to press a ‘yes-button’, were at the start (T1) and at the end (T2) of the second phase. The open label phase randomly mixed with both 10 non-target trials and 10 trials with foils (interference had a duration of 20 weeks; at the end of this period, the final neuropsychological trials) to which the subject had to respond by pressing a ‘no-button’. In the measures were administered (T3). During this open-label phase both groups target trials, the target fruit (a cherry) was shown at a vertical position, whereas were treated with atomoxetine, titrated identically as in the second period. All in the non-target trials the stimulus was absent. In the foil-trials, the stimulus was study personnel, parents and children were blinded to treatment assignment for presented in a horizontal position, which children had to ignore. The stimulus the complete duration of the study. remained on the screen until the child pressed a key. The next stimulus was presented 250 ms after the response. There was no feedback on the response. 168 169 Period 3: Open label, 20 weeks Period 2: Double blind, 8 weeks Period 1: Screening and medication washout Chapter 7 The valid response window was 200 ms to 6000 ms post stimulus onset; trials with responses faster than 200 ms or slower than 6000 ms were replaced by trials of a similar type. Primary outcome measure was the proportion of false alarms on 170 Discontinued (n = 11) due to -adverse event (7) -lack of efficacy (4) Discontinued (n = 3) due to -protocol violation (2) -physician decision (1) interference control were the proportion of false alarms on the non-target trials, the proportion of relevant targets missed, and response time and response time variability of correct responses. ADHD symptomatology T3 (End of open label phase): Previously Placebo (n=35) Assessment + Questionnaires T2(End of double blind phase): Placebo (n=46) Assessment + Questionnaires T1 (Baseline): Placebo (n=49) Assessment + Questionnaires (ADHD-RS)(DuPaul, 1998) and the Conners’ Teacher Rating Scale-Revised: Short Form (CTRS-R:S) (Conners et al., 1998b). The ADHD-RS is a DSM-IV based rating scale administered by a clinician and contains 18 items on inattentive and hyperactive-impulsive symptoms to be scored on a four-point scale in order to assess symptom severity over the past week. The subscales of inattention and hyperactivity-impulsivity only summed the scores of the respective items. If a single item of a subscale was missing, the mean score for all other items in the T3 (End of open label phase): Atomoxetine (n=38) Assessment + Questionnaires T2(End of double blind phase): Atomoxetine (n=43) Assessment + Questionnaires subscale was imputed as the score for the missing item. If more than one item of T1 (Baseline): Atomoxetine (n=48) Assessment + Questionnaires a subscale was missing, the score for the subscale as well as the total score were considered missing. The CTRS-R:S is a 28-item questionnaire to be completed by the child’s teacher in order to assess ADHD-related problem behavior in the school setting. Outcomes on the ADHD-RS and the CTRS-R:S were also reported previously (Harfterkamp et al., 2012). Data analyses Discontinued (n = 4) due to -adverse event (4) Last observation carried forward (LOCF) analyses were conducted (Hamer & Discontinued (n = 5) due to -protocol violation (2) -adverse event (1) -lack of efficacy (1) -parent/caregiver decision (1) Randomized children (N = 97) Excluded (n = 5) based on stated inclusion criteria ADHD symptom improvement was measured using the ADHD Rating Scale Eligible children (N = 102) Baseline Assessment Figure 7.1 Flow diagram for all eligible children suffering from ASD and ADHD symptoms the interference trials. Secondary outcome measures not specifically related to Simpson, 2009), with the requirement of having at least a neuropsychological assessment at T1. In total, 94 out of 97 had a T1 assessment of the response inhibition task and 95 out of 97 children of the interference control task. First, to examine the effect of atomoxetine on both forms of inhibitory control, two analyses 171 Chapter 7 were conducted: a) groups were compared using only the blinded measurements prior (T1) and post treatment (T2); and b) change after first treatment was analysed combining change during the blind phase (T2-T1) for the atomoxetine group and change during the open label phase (T3-T2) for the placebo group. Reaction time and reaction time variability were continuously distributed and were analysed using a) ANCOVA’s, with group as factor, the pre-treatment (T1) as covariate and post-treatment (T2) as dependent measure (Gueorguieva & Krystal, 2004) and b) One-sample T-tests against a test-value of 0 (no change after treatment). Proportions of false alarms and misses showed substantial underdispersion and were therefore analysed using a) negative binominal models with group as factor, the pre-treatment (T1) as covariate and post-treatment (T2) as dependent measure and b) one sample nonparametric tests comparing the median against a testvalue of 0 (no change after treatment). Correlations between congruent measures of both tasks (e.g. reaction time variables of both tasks) were calculated at T1, T2 and T3 to examine the (in)dependence of both inhibitory control domains. Second, inhibitory control measures that significantly changed in response to atomoxetine treatment were used to examine the possible mediating effects of inhibitory control on the change in ADHD symptoms. Parametric (reaction time and reaction time variability) and non-parametric (proportion of false alarms and misses) correlations were calculated between Δ inhibitory control (T2-T1 for the blind group and T3-T2 for the open label group) and Δ ADHD symptoms (T2-T1 for the blind group and T3-T2 for the open label group), in which negative scores reflected improvement (less ADHD symptoms and less erroneous and RESULTS Group characteristics for measures of inhibitory control and ADHD symptom severity are presented in Table 7.2. Effect of atomoxetine on inhibitory control A significant treatment effect was found on the primary outcome measure of response inhibition (proportion false alarms on a Go-No Go paradigm) according to both the blind phase analyses as well as change after first treatment analyses (see Table 7.2). The secondary outcome measure of response inhibition (proportion of misses on the Go-No Go paradigm) only improved during the blind phase. In contrast, the primary outcome measure of interference control (proportion false alarms on interference trials) did not improve after atomoxetine treatment in the blind or open label phase. A secondary measure of interference control (proportion misses) deteriorated after atomoxetine treatment during the blind phase, but not according to change after first treatment analyses. Another secondary measure of interference control (response time) improved according to change after first treatment analyses, but not according to blind phase analyses. Both primary outcome measures (proportion false alarms on a Go-No Go paradigm and proportion false alarms on interference trials) did not correlate significantly (r = .15, p = .17). Except for the proportion of misses in both tasks (r = .38, p < .001), none of the other cognitive or behavioral outcome measures correlated significantly either. faster performances as a results of treatment) 172 173 174 3.39, .07 14.9 (7.3) -2.27 [3.70 - -0.87] 13.83, <.001 -9.39 [-11.52 - -7.26] 25.3 (12.2) n=42/49 -45.57 [-94.91 – 3.77] Note. ATX = Atomoxetine, LOCF = Last observation carried forward. ADHD-RS= ADHD Rating Scale, CTRS-R:S = Conners’ Teacher Rating Scale-Revised: Short Form. 94/97 children had a T1 assessment of the Response inhibition task and 95/97 children of the Interference control task a For 2 children the response inhibition measurement and for 3 children the interference control measurement of T0 was used for T1 because of missing data at T1. b Difference between atomoxetine and placebo post treatment, corrected for pre-treatment. ANCOVAs were used for reaction time (variability) with group as factor, the pre-treatment (T1) as covariate and post-treatment (T2) as dependent measure. Negative binominal models were used for error measures, since these data showed substantial underdispersion, with group as factor, the pre-treatment (T1) as covariate and posttreatment (T2) as dependent measure. c Change after first treatment using one-sample t-test (response time (variability)) or non-parametric tests (error measures) against a test-value of 0 (no change after treatment): double blind phase for the atomoxetine group (T2-T1), open label phase for the placebo group (T3-T2). 13.4 (8.7) 27.5 (12.0) n=39/48 Mediating effect of inhibitory control on ADHD symptomatology A significant treatment effect was found on ADHD symptoms as assessed by the ADHD-RS during the blind and open label phases. For the CTRS-R:S, significant 17.6 (9.0) treatment analyses. Correlations between ADHD-symptoms, error measures and response time (variability) in response inhibition were found to be non-significant (correlations ranging from .00 - .13 for parents and from .02 - .16 for teachers, 15.5 (9.8) 37.6 (9.8) 32.6 (11.0) n=42/48 n=47/49 ADHD symptom improvements were only reported according to change after first all p’s > .10). For interference control, correlations between error measures and response time (variability) and ADHD-symptoms were non-significant either 18.1 (7.5) 38.6 (8.4) n=47/49 (ranging from .04 - .16 for parents and from -.15 - .02 for teachers, all p’s > .10). See also Figures 7.2a and 7.2b for the change of both types of inhibitory control (measured with the proportion of false alarms and proportion misses) in relation 40.7 (7.5) 18.5 (9.3) - ADHD-RS10 11 - CTRS-R:S n=43/48 to the change of ADHD symptoms after both phases. ADHD SYMPTOM SEVERITY -47.22 [-129.60 -18.85] 1081.4 (419.7) 1045.4 (480.7) 985.7 (465.8) 1012.9 (462.3) 931.9 (389.5) 948.1 (458.9) 0.88, .35 - Response time - Response time variability 454.1 (299.6) 459.9 (334.6) 356.0 (263.9) 387.6 (290.3) 372.6 (326.2) 383.2 (276.1) 0.30, .58 0.03 [0.01 – 0.05] 4.88, .03 .06 (.09) .10 (.11) .05 (.06) .06 (.06) .07 (.11) .05 (.06) - Proportion misses -0.02 [-0.05 – 0.01] 0.95, .33 .10 (.12) .10 (.13) .13 (.14) .11 (.14) .12 (.13) .13 (.16) - Proportion false alarms on non-target trials 0.00 [-0.02 – 0.02] 0.10, .75 .05 (.08) .06 (.10) .03 (.07) .04 (.07) .03 (.05) - Proportion false alarms .06 (.10) on interference trials n=46/49 n=42/48 n=47/49 n=43/48 n=48/49 n=47/48 INTERFERENCE CONTROL -2.30 [-17.17 – 12.57] -1.51 [-15.53 – 12.51] 0.00, .99 118.3 (69.4) 133.2 (72.2) 137.7 (79.6) 144.2 (86.5) 120.0 (77.2) 532.7 (144.9) 485.0 (120.7) 525.7 (132.8) 490.5 (104.6) 501.7 (134.8) 491.4 (117.4) 0.07, .80 - Response time - Response time variability 125.1 (64.0) -0.02 [-0.04 - -0.001] 0.00 [-0.01 – 0.01] 3.64, .05 3.83, .05 .06 (.08) .02 (.04) .06 (.07) .01 (.04) .02 (.05) .08 (.11) .04 (.05) .03 (.06) .02 (.05) .06 (.08) .02 (.06) - Proportion false alarms .07 (.10) n=46/49 n=42/48 n=47/49 n=47/48 n=47/49 n=47/48 RESPONSE INHIBITION - Proportion misses LOCF M [95% CI] LOCF F/Wald-χ², p Formerly placebo M (SD) ATX M (SD) Placebo M (SD) ATX M (SD) Placebo M (SD) ATX M (SD) T3: End of open label phase Group Change after first comparison treatmentc blind phaseb T2: End of double blind phase T1: Baselinea Table 7.2 Means and standard deviations of ADHD measures and inhibitory control measures before atomoxetine treatment (T1), after double blind placebo controlled treatment (T2) and after an open label extension (T3) Chapter 7 175 AD open label, proportion misses Interference Control: r = .09, p = .38 for false alarms, r = .16, p = .14 for misses. Interference Control: r = .09, p = .38 for false alarms, r = .16, p = .14 for misses. Response Inhibition: r = .00, p = .99 for false alarms, r = .00, p = .98 for misses. open label, proportion false alarms open label, proportion misses double blind, proportion false alarms double blind, proportion misses Response Inhibition: r = .00, p = .99 for false alarms, r = .00, p = .98 for misses. improved RI, deteriorated ADHD improved RI and ADHD ADHD symptom severity deteriorated RI and ADHD deteriorated RI, improved ADHD Response inhibition (false alarms, misses) Change in response inhibition in relation to change in ADHD symptoms Response inhibition (false alarms, misses) 176 Change after first treatment: Δ Response inhibition, Δ Interference control and Δ double blind, proportion false alarms ADHD symptoms: T2-T1 for the blind double blind, proportion misses phase and T3-T2 for the open label open label, proportion false alarms phase ADHD symptom severity improved IC, deteriorated ADHD improved IC and ADHD ADHD symptom severity deteriorated IC and ADHD deteriorated IC, improved ADHD Change in interference control in relation to change in ADHD symptoms Interference control (false alarms, misses) in ADHD symptoms Figures 7.2a and 7.2b Change in ADHD symptoms and inhibitory control during exposure to atomoxetine in the double blind and open label phases Interference control (false alarms, misses) in ADHD symptoms Chapter 7 DISCUSSION This double-blind placebo controlled study with an open-label extension phase aimed at examining the effects of atomoxetine on two forms of inhibitory control (response inhibition and interference control) in children with both ASD and ADHD, and whether ADHD symptom improvement was mediated by improvement in inhibitory control. The primary outcome measure of response inhibition (number of false alarms on no-go trials) improved as well as the number of misses after atomoxetine treatment; the former both in the blind phase and change after first treatment analyses, the latter only in the change after first treatment analyses (combining blind and open label phase for the atomoxetine and placebo group, respectively). In contrast, the primary outcome measure of interference control (proportion of errors on interference trials) did not improve after atomoxetine treatment and a secondary measure (proportion of misses) actually worsened after atomoxetine treatment during the blind, but not the open label phase. However, performance on the interference control task was somewhat faster after treatment with atomoxetine when change after first treatment analyses were conducted. None of the measures mediated the effect of atomoxetine on ADHD symptom improvement; improvement in ADHD symptom scores and inhibitory control occurred independently from each other. The finding that atomoxetine improves response inhibition is in line with the majority of previous studies documenting response inhibition in children with ADHD (Alderson, Rapport & Kofler, 2007; de Jong et al., 2009; Gau & Shang, 2010; Lijffijt, Kenemans, Verbaten & van Engeland, 2005; Nigg, 1999; Nigg, Blaskey, Huang-Pollock & Rappley, 2002; Oosterlaan, Logan & Sergeant, 1998; Rucklidge & Tannock, 2002; Schachar, Mota, Logan, Tannock & Klim, 2000), response inhibition in adults with ADHD (Chamberlain et al., 2007) and response inhibition in rodents (Bari et al., 2009; Blondeau & Dellu-Hagedorn, 2007; Paterson, Ricciardi, Wetzler & Hanania, 2011; Pattij, Schetters, Schoffelmeer & van Gaalen, 2012; Robinson et al., 2008) although in contrast to a minority (Kuntsi, Oosterlaan, & Stevenson, 2001). Our study adds to these findings by 177 Chapter 7 illustrating that response inhibition can also be improved in patients with ASD & Taylor, 2005), whereas for interference control, a reduced activitity of the frontal- and comorbid ADHD, tentatively suggesting that the origins of response inhibition striatal cortex (including the cingulate cortex) as well as parietal cortex has been problems in these comorbid patients may be of similar nature as those observed reported in ADHD (Bunge et al., 2002; Lee, 2010; Vaidya et al., 2005). Atomoxetine in ADHD-only patients. This hypothesis is further supported by the improvements appears to mainly exert its effects on the prefrontal cortex underlying response of ADHD symptoms as a result of treatment with atomoxetine in this patient inhibition, and not on the parietal brain regions also required for interference group(Harfterkamp et al., 2012) as well as in previous (mostly open-label) studies control in patients with ASD and ADHD. In other words, the enhanced frontal- (Arnold et al., 2006; Charnsil, 2011; Fernández-Jaén et al., 2012; Hazell, 2007; striatal functioning via the selective norepinephrine re-uptake inhibitor may be Jou et al., 2005; Posey et al., 2006; Troost et al., 2006; Zeiner et al., 2011). All in sufficient to improve response inhibition, but insufficient to improve interference all, atomoxetine appears to improve both response inhibition problems as well suppression because parietal-temporal contributions are necessary (Bunge et as ADHD symptoms in patients with ASD and ADHD, suggesting this to be an al., 2002; Casey et al., 2000). Further functional imaging studies are required to effective method of treatment in these patients. confirm or refute this hypothesis. In contrast, the effect of atomoxetine on interference control was absent, An intriguing finding reported here was the lack of association between which runs counter to previous studies examining the effect of atomoxetine on ADHD symptom improvement and improvements in inhibitory control. The interference control in ADHD patients (Grodzinsky & Diamond 1992; Spencer absence/presence of such a relationship is a relatively neglected topic in studies et al., 1998; Yang et al., 2011), but is in line with others (Faraone, Biederman, reporting on pharmaceutical effects on cognitive and symptom data (Boonstra, et al., 2005; Schwartz & Verhaeghen, 2008; Spencer et al., 2006; van Mourik, Kooij, Oosterlaan, Sergeant & Buitelaar, 2005; Fernández-Jaén et al., 2012; Oosterlaan & Sergeant, 2005).The differential effect of atomoxetine on response Turner, Clark, Dowson, Robbins & Sahakian, 2004), yet is of great relevance in inhibition on the one hand and interference control on the other hand, is not understanding the mechanisms of action of a pharmacological agent. Some surprising given that these functions do not necessarily correlate with each other, previous studies on the effects of atomoxetine on ADHD symptoms and inhibitory which was the case in our study as well as previous studies (Barkley, Grodzinsky control did report on this relationship, and found no connection between & DuPaul, 1992; Scheres et al., 2003; van der Oord, Geurts, Prins, Emmelkamp & improvement of one and the other (Gualtieri & Johnson, 2008; Spencer et al., Oosterlaan, 2012). The neural substrates of response inhibitionand interference 2006; Vaughn et al., 2011). Similarly, independent effects of methylphenidate on control in ADHD are not yet fully understood. It has been suggested that alteration ADHD symptoms and associated cognitive deficits have been found (Epstein et in the neural basis of response inhibition and interference control in childhood al., 2006; Scheres et al., 2003), suggesting this is not a rare phenomenon or a ADHD are characterized by distinct patterns of functional abnormality (Bunge, finding only reported in atomoxetine treatment studies. It has been suggested that Hazeltine, Scanlon, Rosen & Gabrieli, 2002; Vaidya et al., 2005). The functional cognitive deficits in psychiatric disorders may act as epiphenomena: related to activity associated with response inhibition in ADHD is a decreased activation of the same etiological underpinnings as the disease symptoms, but not mediating the prefrontal and striatal brain regions (including the inferior frontal gyrus and between both (Kendler & Neale, 2010), it is still unclear for which functions and cingulate cortex) (Aron & Poldrack, 2005; Durston, 2003; Durston, Mulder, Casey, for which disorders this is the case. However, it may suggest that predicting or Ziermans & van Engeland, 2006; Rubia et al., 1999; Rubia, Smith, Brammer, Toone monitoring treatment response using cognitive tests may not be of clinical utility, 178 179 Chapter 7 at least in the case of atomoxetine treatment in children with ASD and ADHD using inhibitory control tasks. Several limitations warrant consideration. First, we did not include an unaffected control group and can therefore not be certain of a baseline deficit in inhibitory control. However, since all children were clinically diagnosed with ADHD, at least a proportion was likely to have inhibitory control difficulties (Barkley, 1997; Pennington, 1996), and our results showed that for the overall group room for improvement was present given the significant effect of atomoxetine treatment on response inhibition. Second, we used only two tests to measure inhibitory control, perhaps the use of a more comprehensive task battery could have led to more firm conclusions on the efficacy of atomoxetine on inhibitory control in children affected with both disorders. Third, our study sample had relatively few adolescents, female subjects, and children with IQ’s in the lower range, making findings possibly less accountable for these groups. In sum, this is the first double blind, placebo controlled study to demonstrate selective beneficial effects of atomoxetine on motor response inhibition but not interference control in children with both ASD and ADHD symptoms. This cognitive benefit occurs independently from improvements in ADHD-symptomatology, which may suggest that distinct pathogenic pathways play a role in clinical symptoms and cognitive dysfunctions. Increased understanding of the effects of medical treatment on both levels of functioning can help the development of personalized medicine; medication with a higher probability of desired outcomes and reduced side effects. 180 181 General discussion 182 ASD and ADHD are frequently co-occurring neurodevelopmental disorders that are rather heterogeneous in symptom presentation and underlying etiological factors. ASD is characterized by impaired social interaction, impaired verbal and nonverbal communication, as well as restricted and repetitive behavior and interests, while ADHD is characterized by severe inattention, hyperactivity and impulsivity (American Psychiatric Association, 2013). Most estimates for the presence of ADHD among patients with ASD fall within the range of 30% to 80%, whereas the presence of ASD is estimated in 20% to 50% of the patients with ADHD (e.g. Ames & White, 2011; Leyfer et al., 2006; for review see Rommelse et al., 2010; Ronald et al., 2008). ASD and ADHD are both typified by cognitive impairments (for review see Rommelse et al., 2011). The overall objective of this thesis was to examine shared and unique behavioral and cognitive profiles in ASD and ADHD, and whether these profiles provided more insight into the shared and unique etiology of ASD and ADHD. The disclosure of shared and unique underlying mechanisms of ASD and ADHD is complicated by the heterogeneity across diagnostic categories in the DSM classification system, for group comparisons based on DSM classifications reflect symptoms rather than causes (e.g. Lord & Jones, 2012; Miller, 2010). In addition, approaches based on DSM-cut-offs overlook the evidence that the general population (i.e. under the clinical cut-off) is also characterized by behavioral, cognitive and genetic variance (e.g. Barnett et al., 2009; Constantino, 2011; Fair et al., 2012; Plomin et al., 2009; Robinson, Koenen et al., 2011). In line with these insights, this dissertation used dimensional measures of ASD and ADHD symptoms, which more accurately reflect the continuously distributed and multifaceted nature of behavioral symptoms and cognitive profiles of both disorders within the population. To empirically dissect heterogeneity and define more homogeneous disease profiles, an empirical bottom-up approach (i.e. latent class analyses (LCA) (McCutcheon, 1987)) was applied. In sum, the approach of this thesis had the following key characteristics: 185 Chapter 8 I) Assess ASD and ADHD symptoms in parallel. were largely linear. That is, for any behavioral and cognitive outcomes that showed II) Examine the relationship between behavior and cognition. significant associations with the ASD or ADHD traits, the non-symptomatic ends III)Apply dimensional measurements by focusing on both the lower and of the trait continua were associated with fewer behavioral problems and better higher end of the ASD and ADHD trait continua and integrating data from cognitive performances than the symptomatic ends. This provided support population-based and clinic-based samples. to the assumption that ASD and ADHD are both extreme ends of continuous IV)Identify subgroups that are homogeneous at the behavioral or cognitive level by using latent class analyses. traits that cover quantitative rather than qualitative differences. Future studies on the correlates of the non-symptomatic ends of both continua may deepen This general discussion summarizes and discusses key findings from all chapters, our understanding of protective mechanisms underlying superior behavioral and points out limitations, suggests recommendations for future research, and closes cognitive functioning. with some clinical implications. The relationship between behavior and cognition: reduced heterogeneity Summary on the behavioral level The continuum of ASD and ADHD traits In chapter 3, LCA on ASD and ADHD symptom data were conducted in order Chapter 2 focused on normally distributed ASD and ADHD traits in a populationbased sample of 378 children between 6 and13 years of age. ASD and ADHD are thought to exist on a continuum, where diagnosis simply reflects the symptomatic end of a normal distribution of quantitative traits in the general population. This study was conducted in order to find support for or evidence against the hypothesis that the lower extreme ends of the trait continua are associated with favorable cognitive and behavioral outcomes. Alternatively, extreme deviations in either direction on the continua may be pathological. That is, being at the lowest risk for ASD or ADHD may also come with specific disadvantages. For example, very low levels of restrictive and repetitive behaviors may lead to difficulties keeping a daily routine, and being highly reflective instead of impulsive may lead to inertia. We examined if the association of ASD and ADHD traits with cognitive and other behavioral measures was linear or curvilinear. Findings indicated that the non-symptomatic ends of the ASD and ADHD trait distributions indeed represented some superior cognitive (for ADHD) and behavioral (for ASD and ADHD) functioning. The associations between the ASD and ADHD traits on the one hand and behavioral problems and cognitive functioning on the other hand 186 to find support for or evidence against the hypothesis that ASD and ADHD are different manifestations of one overarching disorder. We applied LCA on Social Communication Questionnaire (SCQ) and Conners’ Parent Rating Scale (CPRSR:L) data of 644 children between 5 and 17 years of age from both a population and clinic-based sample. Classes were compared for comorbid symptoms and cognitive profiles of motor speed and variability, executive functioning, attention, emotion recognition and visual spatial functioning. LCA identified three patient classes that could be distinguished from two normal classes: one class with ADHD symptoms only, one class with clinical levels of ADHD but also clinically elevated levels of ASD symptoms (ADHD[+ASD]), and one class with clinically high levels of ASD symptoms but also clinically elevated levels of ADHD symptoms (ASD[+ADHD]). As hypothesized, no class with exclusive ASD symptoms was revealed; all children who expressed ASD behavior also presented the less severe ‘precursor’ of ADHD behavior. These findings gave support for the gradient overarching disorder hypothesis, which states that ADHD may best be seen as a milder, less severe subtype within the ASD spectrum. In addition, the cognitive profiles partially supported the gradient overarching disorder hypothesis as 187 Chapter 8 well. That is, the cognitive functioning in the ADHD-only class could overall be In chapter 5, we used measures that provide greater resolution of the scores in considered at an intermediate level, but qualitative differences were also observed, the lower end of the ASD and ADHD trait distributions, subjected these to LCA, with working memory deficits being more pronounced in both primarily ADHD and examined how subgroups differed in terms of cognitive functioning. Analyses classes compared to the primarily ASD class, and a detail-focused cognitive style revealed five classes; three mostly quantitatively differing concordant ASD-ADHD in visual pattern recognition in the ASD(+ADHD) class only, despite the fact that classes with either low, medium or high scores on both traits (77.5%), which both classes showed clinically elevated ASD and ADHD symptoms. Thus, we indicated that ASD and ADHD traits are usually also strongly related in the general concluded that different ASD–ADHD comorbid subtypes exist, with quantitative population. Two discordant ASD-ADHD classes presented with either more ASD overlap, but also qualitative difference in cognitive deficits. This has clinical symptoms than ADHD symptoms or vice versa (22.5%) were characterized by relevance, since these children may respond differently to medication treatment, differential visual spatial functioning. These findings across the non-symptomatic social skills therapy and behavioral therapy. end of the ASD and ADHD traits closely resembled those across the clinical ASD- In chapter 4, the 8 to 17 year old children from the Normal, ADHD-only, ADHD classes described in chapter 3, which suggested that heterogeneity in ASD ADHD(+ASD) and ASD(+ADHD) classes described in chapter 3were examined and ADHD may be rooted in heterogeneity present in the general population. with regard to their motor timing abilities. Motor timing abilities are frequently reported to be affected in children with ADHD, and deficits in time processing may The relationship between behavior and cognition: reduced heterogeneity play an important role by modulating primary symptoms. For example, attention, on the cognitive level language and inhibition are associated with time processing, as these functions Chapter 6 examined the association between behavior and cognition in the are characterized by specific temporal patterns. Results indicated that motor reversed direction, subjecting cognitive measures to LCA, a rather novel approach. timing accuracy was more affected in children with ADHD with comorbid ASD Given that reduced heterogeneity on the behavioral level resulted in informative symptoms (ADHD[+ASD]) compared to the children in the ADHD-only class. cognitive profiles in chapter 3, segmenting on the basis of cognitiveperformances This finding led to the conclusion that only patients with more severe behavioral may be a useful complementary approach in detecting shared and unique symptoms show motor timing deficiencies. However, this could not merely be mechanisms underlying ASD and ADHD. That is, cognitive functioning can be explained by high ADHD severity with ASD playing no role, as ADHD symptom measured more objectively than clinical symptoms, and is possibly more closely severity in the pure ADHD-class and the ASD(+ADHD) class was highly similar, linked to the neurobiological underpinnings than behavior (Gottesman & Gould, with the former class showing no motor timing deficits. This was well in line 2003). Based on the findings described in chapter 3, it was hypothesized that with the gradient overarching disorder hypothesis described in chapter 3, and cognitive subtypes might be identified with a) superior visual pattern recognition underlined the relevance of a reduction of the behavioral heterogeneity present in skills and inferior emotion recognition abilities that was most strongly linked to ASD; DSM-defined ASD and ADHD group comparisons. and b) inferior visual pattern recognition skills and normal emotion recognition In chapter 5, the focus shifted back to the population-based sample only. abilities that was most strongly linked to ADHD. Contrary to these hypotheses, the This sample was mostly grouped together through LCA into ‘Normal’ classes on main finding was that LCA in the clinic and population-based samples revealed a the basis of clinical (thus skewed) ASD and ADHD measures used in chapter 3. similar four class solution typified by qualitatively different speed-accuracy trade- 188 189 Chapter 8 offs instead of domain specific strengths/weaknesses. Furthermore, in the clinic- that cognitive deficits in psychiatric disorders may act as epiphenomena: related based sample, these speed-accuracy trade-off patterns were strongly linked to to the same etiological underpinnings as the disease symptoms, but not mediating between-class differences in ASD and ADHD (and comorbid) symptom severity. between both. This is clinically relevant, since it may suggest that predicting or More specifically, the cognitive profile that overall favored accuracy over speed monitoring treatment response using cognitive tests is not of clinical utility in was associated with the lowest amount of ASD and ADHD symptoms, while the atomoxetine treatment in children with ASD and ADHD using inhibitory control profile characterized by slow and inaccurate performance across the cognitive tasks. tasks was associated with the highest amount of ASD and ADHD symptoms. These associations between cognitive functioning and behavioral symptoms Discussion were much weaker in the population-based sample, which was not due to a On the relationship between ASD and ADHD restriction of variance in symptom measures, since normally distributed ASD and ADHD trait measures were additionally used. The absence of cognitive profiles with domain specific strengths/weaknesses contrasts with the idea of multiple cognitively different developmental pathways to ASD and ADHD and the cognitive profiles detected in chapter 3. This finding suggested that cognitive functioning only predicts behavior when other risk or protective factors are present. Further investigation of the relationship between behavior and cognition: clinical trial data The clinical trial described in chapter 7 examined whether a pharmacologic intervention in the noradrenergic system hypothesized to improve symptoms of ADHD would also improve cognitive impairments in children with both ASD and ADHD. This double blind, placebo controlled study demonstrated that children improving in response inhibition were not necessarily the ones also improving in ADHD symptoms, since none of the cognitive measures mediated the effect of atomoxetine on ADHD symptom improvement. That is, although both response inhibition and ADHD symptoms improved under the influence of atomoxetine, improvement in response inhibition did not correlate with improvement in ADHD symptoms. This lack of association between ADHD symptom improvement and improvements in inhibitory controlmay suggest that distinct pathogenic pathways play a role in clinical symptoms and cognitive dysfunctions. It has been suggested 190 With respect to the relationship between ASD and ADHD, the most fundamental issues are at the nosological level: Are ASD and ADHD distinct disorders, or do they reflect an arbitrary division of a single syndrome (Neale & Kendler, 1995)? Multiple findings appeared to suggest that ASD and ADHD are alternate expressions of a single underlying dimension of liability, as described by Banaschewski et al. (2007) and Rommelse et al. (2010). That is, ASD and ADHD seemed to present with shared etiological substrates, in which ADHD may best be seen as a milder, less severe subtype within the ASD spectrum (chapters 3, 4 and 5). In short, the most important findings in favor of this hypothesis were: a) the presence of a latent class with children that presented with ADHD symptoms without comorbid ASD symptoms, in contrast to the absence of a latent class with children who present with ASD without comorbid ADHD symptoms; b) the latent class with only ADHD-symptoms presented with behavioral and cognitive difficulties which were on an intermediate level when compared to the class without behavioral problems and the class with ADHD symptoms and comorbid ASD symptoms; c) mostly quantitatively differing concordant ASD-ADHD classes with either low, medium or high scores on both traits were found in both clinic-based samples and across the population-based sample. 191 Chapter 8 Considering ADHD as a less severe subtype within the ASD spectrum calls for liability model. Nonetheless, no definite conclusions can be drawn since a study a broader view on developmental disorders, as will be discussed in the future design with repeated measures that monitors behavior over time is warranted directions below. In contrast, another finding was more in favor of a distinct to distinguish between a shared underlying liability and reciprocal causation. disorder hypothesis for ASD and ADHD: Such a research design was applied in the Avon Longitudinal Study of Parents a) some specificity of cognitive deficits with regard to the visual spatial and Children (ALSPAC) study conducted by St.Pourcain and colleagues (2011). processingstyle (i.e. detail-focused or not), working memory and emotion They stated that the clinical presentation of both ASD and ADHD is influenced by recognition across latent ASD-ADHD classes (chapters 3 and 5). age, with ASD symptoms being more stable compared to ADHD symptoms. As The difference in visual spatial functioning between ASD and ADHD reflected a changes in symptom severity for ASD and ADHD varied, both traits were strongly double dissociation, reflective of a qualitatively different cognitive deficit. This was but not reciprocally interlinked. That is, the majority of children with a persistent in line with recent studies that reported ADHD-subgroups with unique cognitive ADHD symptomatology also showed persistent ASD deficits but not vice versa, and/or comorbid behavioral profiles (Fair et al., 2012; Pauli-Pott, Dalir, Mingebach, suggesting that a complex and variable relationship exists between the ASD and Roller & Becker, 2013). These findings led the authors to suggest that the clinical ADHD traits (St. Pourcain et al., 2011). phenotype of ADHD may be rooted in multiple distinct cognitive subgroups and/ or developmental pathways. If true, the identification and description of such ASD-ADHD comorbidity, we can conclude that some substrates of ASD and subgroups may help clinical practice to meet the specific deficits of these children ADHD may be shared between both disorders, while others may be unique. by more tailored diagnostic and treatment procedures within a personalized The gradient overarching disorder hypothesis, in which ASD and ADHD are (medicine) framework. quantitatively different from one and another and in which ADHD may best be Finally, a third potential model of comorbidity that cannot be rejected seen as a milder subtype within the ASD spectrum, seems a reasonable model. on the basis of the findings in this thesis is that ASD-ADHD comorbidity is due That is, most findings are well in line with this hypothesis, except for visual spatial to reciprocal causation (Banaschewski et al., 2007). That is, ADHD may increase processing styles, which may indicate why some children develop ADHD despite the risk for ASD (‘domino effect’), which can explain the presence of pure ADHD their enhanced risk for ASD. Future research on causal models for ASD, ADHD as well as comorbid ASD-ADHD. However, the reverse pattern, i.e. that ASD may and ASD-ADHD comorbidity may benefit from longitudinal research designs, as increase the risk of ADHD, would suggest the presence of pure ASD, which we did patterns of association between ASD and ADHD symptoms change over time not find. The nature of our samples might have prevented the detection of pure (St. Pourcain et al., 2011). In addition, future research may wish to use not only ASD, as ASD without comorbid ADHD symptoms may be underrepresented in behavioral and cognitive measures, but also account for the influence of potential clinical samples and may also be relatively rare in population samples. Therefore, environmental and genetic risk or protective factors in order to resolve this very large population-based samples might be best to examine whether children nosological uncertainty (Jaffee & Price, 2007). Based on the current findings, and regardless of the exact cause for who present with ASD without comorbid ADHD exist, but their rarity by itself may suggest that a reciprocal risk is an unlikely route. Rather, our data suggest that if ASD is present, ADHD seems present as well, which is more in line with the shared 192 193 Chapter 8 On the relationship between behavior and cognition Cognitive measures are used frequently in the assessment of ASD and ADHD. These measures are also referred to as intermediate phenotypes, useful indicators in detecting etiologically more homogeneous subgroups of patients (Gottesman & Gould, 2003). Intermediate phenotypes form a causal link between genes and behavioral symptoms, more closely linked to the genes in action in ASD and ADHD, and more objectively measured than behavior (Kendler & Neale, 2010). Characteristic of intermediate phenotypes is that they are heritable, associated with the disorder, state independent and present in non-affected family members of patients (Walters & Owen, 2007). A growing body of evidence strongly suggests that, although several cognitive domains can impact on symptom severity, none of them is necessary or sufficient in causing the developmental disorders (see for review Coghill, Seth, et al., 2013). Kendler and Neale (2010) aimed to clarify the relationship between behavior and cognition through a conceptual analysis of the intermediate phenotype construct. Among the concerns discussed were that a) no distinction could be made between a liability-index model and a mediational model for intermediate phenotypes, b) the association between an intermediate phenotype and a psychiatric disorder could either be unidirectional and/or bidirectional, and c) intermediate phenotypes may reflect both environmental and/or genetic risk factors for the development of a disorder. For those reasons, a relevant and powerful strategy to study the relationship between behavior and cognition via intermediate phenotypes would be a longitudinal research design with multiple measurements of behavior and cognition over time. Such designs have the capacity to clarify the causal relationships between intermediate phenotypes and psychiatric disorders. The complexities in the relationship between behavior and cognition only when additional risk factors are present. The associations between cognitive measures and symptom severity were weak or absent in the population-based sample, providing little evidence in support of the hypothesis that ASD and ADHD traits share cognitive underpinnings in the general population. In contrast, the associations between cognition and symptoms were related to symptom severity in the clinic-based sample, suggesting that the measures used did uncover cognitive mechanisms underlying ASD and ADHD. Such a missing link between symptom severity and neurocognitive functioning in the general population was also reported in another recent population-based study (Rajendran, Rindskopf, et al., 2013), while observed in a clinic-based sample (Rajendran, Trampush, et al., 2013). In the studies described in this thesis, it is unlikely that a restriction of variance in symptom measures can explain this difference in findings between both samples, since the findings pertained also to normally distributed ASD and ADHD trait measures. Rather, it seems likely that the relationship between cognitive functioning and symptom severity may be more complex than expected. Coghill, Hayward and colleagues (2013) examined this complex relationship further through a prospective study of the association between compensatory improvements in executive neuropsychological functioning and clinical symptom improvement. Their findings suggested that cognitive deficits and behavioral symptoms are not linearly related, such as depicted in Figure 8.1, which is a slightly adapted version of a traditional causal model for ASD and ADHD described by Coghill, Hayward, et al. (2013). Rather, the relative lack of association between clinical and cognitive changes over time may suggest that clinical and cognitive aspects of ASD and ADHD make independent contributions to overall impairment; see the slightly adapted version of a potential alternative causal model depicted in Figure 8.2. are illustrated by our data. The treatment study described in chapter 7 indicated that none of the inhibitory control measures was associated with ADHD symptom improvement. In addition, the findings described in chapter 6 suggest that the relationship between behavior and cognition may be stronger and can be detected 194 195 Chapter 8 Figure 8.1 Traditional causal model for ASD and ADHD, inspired by Coghill, Hayward, et al. (2013) addition, McAuley, Crosbie, Charach & Schachar (2013) also concluded that cognitive performances (i.e. response inhibition) in ADHD improved with age regardless of changes in ADHD symptoms and impairment. Thus, cognitive Environment deficits were state-independent, present irrespective of changes in the behavioral symptoms. Third, non-affected siblings of ASD and/or ADHD affected children present with resembling cognitive difficulties, as was also outlined by Oerlemans Brain structure and function Genes ASD/ADHD symptoms Cognition Impairment and colleagues (2013). Here, the cognitive performances of unaffected siblings were overall at an intermediate level, performing somewhat worse than healthy control children and better than their ASD affected siblings. These findings add Figure 8.2 Potential alternative causal model for ASD and ADHD, inspired by Coghill, Hayward, et al. (2013) Environment up to the idea that distinct pathogenic factors underlie behavioral symptoms on the one hand, and specific cognitive dysfunctions on the other hand. Although multiple findings are in line with this alternative causal model, another possible explanation may be that the extent to which cognitive tasks accurately measure daily life difficulties (e.g. difficulties in time management and self motivation for ADHD, difficulties in social reciprocity and changes in daily routines for ASD) may Cognition Genes Brain structure and function be limited (Barkley & Fisher, 2011; Knouse, Barkley & Murphy, 2013). Rather, Impairment ASD/ADHD symptoms research indicated that the recently developed questionnaire Barkley Deficits in Executive Functioning Scale (BDEFS) correlated more closely to daily difficulties in ADHD than did cognitive deficits (Barkley, 2011). In any case, the uncertainty and complexity of the relationship between symptoms and cognition calls for Multiple findings are in line with this latter model. First, the change in performances on the specific cognitive measures used in chapter 7 (i.e. inhibitory control, which is a robustly validated neurocognitive deficit in ASD and ADHD (for meta-analyses see Geurts, van den Bergh & Ruzzano, epub ahead of print; for review see Rommelse et al., 2011) appeared to be unrelated to change in ADHDsymptomatology. Second, the systematic review on the cognitive predictors of persistence of ADHD conducted by van Lieshout, Luman, Buitelaar, Rommelse & Oosterlaan (2013) did not provide evidence to support the hypothesis that either automatically controlled or more consciously controlled cognitive functions differentiated between persistence and remittance of ADHD symptoms. In 196 cautiousness in the clinical neurocognitive field and the use of intermediate phenotypes. The studies described in this thesis tentatively discussed potential relations between the cognitive performance patterns in ASD and ADHD, and their neural basis. The more generic cognitive profiles detected in chapter 6 are in line with theories of more diffuse abnormal cortical connectivity and synchrony as well as more widespread neural abnormalities in ASD and ADHD (Rudie et al., 2013; Wang et al., 2009; Wass, 2011). Long-range cortical under-connectivity and local (sub)cortical over-connectivity are found in the majority of studies in ASD; an increased local connectivity was found in some but not all studies across children 197 Chapter 8 with ADHD (e.g. Bush et al., 2005; for review see de la Fuente et al., 2013; Silk et brain on the other hand, a humble attitude towards our current knowledge of the al., 2005; Vance et al., 2007; for review see Vissers et al., 2012). This difference relationship between behavior and the brain seems appropriate. between ASD and ADHD seems in accordance with the double dissociation in visual-spatial functioning described above. In short, these findings indicated that On the use of a dimensional approach children displaying more ASD traits than ADHD traits presented with a visual Previous studies have shown that ASD and ADHD exist on a continuum that perceptual processing style that facilitated local rather than global processing, extends from the symptomatic end of ASD and ADHD affected populations to while the opposite pattern was seen in children that displayed more ADHD traits the non-symptomatic end without ASD and ADHD symptoms (Constantino, 2011; than ASD traits. Although this dissociation may pinpoint towards differences in Fair et al., 2012; Plomin et al., 2009; Robinson, Koenen, et al., 2011). In chapter 2, connectivity (i.e. favoring local processing in ASD, and favoring global processing the lower ends of the ASD and ADHD trait continua appeared to represent largely in ADHD), it is still too early to draw conclusions. That is, some inconsistency in quantitative rather than qualitative differences. That is, the non-symptomatic ends findings may also be due to the multiple definitions that are used for ‘long range’ of the ASD and ADHD trait continua were associated with lower levels of behavioral and ‘local’ connectivity (Vissers et al., 2012). Future research therefore warrants problems on gold-standard measures for internalizing and externalizing behavior, an explicit definition of ‘long-range’ and ‘local’ connectivity, in order to develop a and some advantages in cognitive performances for the ADHD continuum. In better understanding of brain connectivity in ASD and ADHD. In general, future chapter 5, we did demonstrate that many of the findings in non-clinical ASD- research should aim at relating cognitive processes to the diffuse abnormal ADHD classes were in line with the findings across the clinical ASD-ADHD classes cortical connectivity and synchrony and widespread neural abnormalities described in chapter 3. This indicated that variance in ASD and ADHD is rooted seen in ASD and ADHD. This can for example be achieved through functional in variance in the non-symptomatic end of the trait distributions. Furthermore, neuroimaging during cognitive task completion. The cognitive performances in chapter 6, resembling four class solutions typified by qualitatively different mapped on connectivity measures may provide more precise knowledge on the speed-accuracy trade-offs emerged in both the population-based and clinic- relationship between activity in certain brain areas and specific cognitive -and based samples. This suggests that the cognitive profiles disclosed through a ultimately behavioral- measures. bottom-up approach are generic, i.e. not only relevant for neurodevelopmental Although structural and functional alterations in widespread brain disorders but also for normal development. These conclusions are in line with the regions and their connections may be central in ASD and ADHD pathophysiology perception of ASD and ADHD as quantitative traits rather than as categorically (for reviews see de la Fuente et al., 2013; Vissers et al., 2012), still little is known defined conditions. And clearly, some of the heterogeneity in neurodevelopmental about their exact relevance for the expression of ASD and ADHD behavior. disorders is rooted in heterogeneity present in the general population. Stronger behavioral links are necessary to deepen our understanding of the actual importance and influence of abnormalities in brain connectivity to the heterogeneity between affected and normal populations has implications for manifestation of ASD and ADHD. Given the complexity of the relationships not only diagnostic classification and detection of underlying mechanism of between behavior and cognition on the one hand, and between cognition and the neurodevelopmental disorders, but also for public health efforts to identify and As described by Constantino (2011), this finding of continuity of support affected children. Public health monitoring can be more accurate if it 198 199 Chapter 8 is not limited by clinical diagnoses and cut-offs. Ideally, everyone should be was concluded that ASD and ADHD were typified by some uniqueness in their characterized by quantitative risk scores on an established amount and selection cognitive profiles (regarding emotion recognition and visual spatial functioning). of trait continua. Research thus far indicated that continuum scores are stable over The study described in chapter 6, where a cognition-based multivariate bottom- time for ASD, and of moderate stability for ADHD across the general population up classification was conducted, resulted in cross-domain generic cognitive (Constantino et al., 2009; Lakes, Swanson & Riggs, 2012; Robinson, Munir, et al., factors rather than domain specific strengths/weaknesses. Thus, it appears that 2011). The use of continuous risk scores may potentially lead to a public health results partially depend on the measures used. It is therefore recommended that care model that focuses on prevention rather than solely on treating patients, not only symptom counts, but also aspects such as its context-dependency and as elegantly outlined by Plomin and colleagues (2009). In chapter 2, it was longitudinal clinical course are explored in the development of valid symptom concluded that the non-symptomatic ends of the ASD and ADHD trait continua dimensions for bottom-up classification (Uher & Rutter, 2012). In addition to represented some superior functioning, possibly reflective of not only low risk, but these measures, future research may also wish to conduct gene-based and/or potentially even of resilience for developmental disorders. Increased insight into functional neuroimaging-based multivariate bottom-up classifications. the correlates of this low risk / resilient end of the trait distribution may deepen our insights in protective factors. As described in chapter 6, potential protective care with insights in protective factors and resilience, opening doors towards factors for the development of behavioral problem may be a cognitive profile that prevention rather than cure. A new taxonomy however does require an established favors accuracy over speed, a higher intelligence, and a somewhat older age amount and type of measures to accurately ascertain their relative contributions (maturation). to neurodevelopmental problems. See also the future directions below for the It is tempting to speculate about the usefulness of a bottom-up promising Research Domain Criteria project (RDoC) approach developed by the classification instead of DSM-based classification system for clinical practice. As National Institute of Mental Health (NIMH)(http://www.nimh.nih.gov/research- described by Uher & Rutter (2012), symptom dimensions can only be considered priorities/rdoc/index.shtml). To conclude, a dimensional approach may provide public health as alternatives to diagnosis, after a systematic approach to the selection of symptom variables tackled the hard questions of how many and which dimensions Key findings are really needed. That is, dimensions have the potential to enhance the validity • Bottom-up subtyping into behaviorally homogeneous groups reveals of classification, but the sensitivity to input methodology and lack of consensus associations between behavior and cognition that easily remain undisclosed suggest that it is still too soon to replace categories with dimensions. To be of in DSM-defined heterogeneous subgroups. clinical utility, findings should not be affected by differences in for example the • Findings in line with the gradient overarching disorder hypothesis are that samples or measures used across studies. For instance, despite the fact that ASD and ADHD do differ quantitatively but not qualitatively from one and both studies were conducted across the same populations, the conclusions from another in baseline motor speed and reaction time variability, executive the complementary behavioral and cognitive empirical bottom-up approaches functioning, attention and emotion recognition. For these measures, ADHD in the chapters 3 and 6 may appear contradictory at first sight. In chapter 3, may best be seen as a milder subtype within the ASD spectrum. where a behavior-based multivariate bottom-up classification was conducted, it 200 201 Chapter 8 • • • • • • • Findings in contrast to the gradient overarching disorder hypothesis are corrected for the influence of age, for example by calculating residuals, by that visual spatial processing styles do differ qualitatively between ASD and covarying for age and by using age-corrected subscale scores of behavioral ADHD, with a detail focused visual spatial processing style typical for ASD but questionnaires. Notwithstanding, the optimal approach to tackle a potential age- not ADHD. This may indicate why some children develop ADHD despite their effect would be the use of a longitudinal research design. enhanced risk for ASD. It is less likely to observe pure ASD in clinical practice; generally ASD is ASD and/or ADHD, in both the clinic-based and population-based sample. This presented with co-occurring ADHD symptoms. corresponds with the referral bias in clinical practice, and the upper extreme ASD Motor timing difficulties, possibly interacting with primary ASD and ADHD and ADHD traits being more easily recognized in boys than in girls (Kramer et al., symptoms, cannot merely be explained by ADHD only, as ADHD symptom 2008). As sex is inherently confounded with ASD and ADHD, the influence of sex severity in the pure ADHD-class and the ASD(+ADHD) class was highly could not be separated from the effects of the ASD and ADHD symptomatology. similar, with the former class showing no motor timing deficits. In addition, the sex-differences at the upper extreme end of phenotypic traits An overall cognitive slower and more inaccurate performance is an aspecific are also observed in the general population. This suggests that sex differences but strong predictor of psychiatric dysfunctioning, which calls into question in clinical referral patterns and diagnoses of ASD and ADHD are not based on the use of extensive and specific cognitive batteries that are now often used a clinical bias, but rather reflect a true predisposition in males (Neuman et al., for research and clinical purposes. 2005). The cognitive profiles observed in the clinic-based sample resemble those across the population-based sample, which may indicate that part of the information on behavioral problems. This did not affect comparisons between cognitive heterogeneity in ASD and ADHD is nested in normal variation. classes and between studies, but in contrast to clinical interviews, questionnaires Cognitive functioning may only predict behavioral problems when other risk do not allow for further probing or explanation of questions. In addition, factors are present. In every day life, favoring speed over accuracy or the information from multiple informants instead of parents-only may have provided other way around may be equally adaptive. a more generalized behavioral profile. Information on children by means of Although both response inhibition and ADHD symptoms improved under questionnaires completed by teachers was collected, and can be used for further the influence of atomoxetine, improvement in response inhibition did not analyses. Second, boys were overrepresented in the classes with higher levels of Third, questionnaires completed by parents were used to collect correlate with improvement in ADHD symptoms. Limitations These studies come with some limitations and considerations. First, the children varied widely in age (5-17 years). Since age has a strong effect on both cognitive performances as well as type of behavioral symptoms presented, the influence of age may have affected between class differences. However, analyses were 202 203 Chapter 8 Future directions unintended consequences of reifying the current diagnoses that probably do not Development of diagnostic classification schemes mirror nature. Such a process may help develop and adjust the classification When comparing the current DSM-5 with its antecedents (i.e. DSM-III, DSM-III-R, DSM-IV and DSM-IV-TR), some progress can be observed. In earlier psychiatric classification schemes, many of the diagnoses included hierarchical exclusionary rules such that certain diagnoses could not be assigned if the symptoms occurred during the course of another disorder that occupied a higher level in the hierarchy (e.g. ADHD was excluded in the presence of ASD). These exclusion rules were later seen as problematic because these were not empirically based and made the study of lower-ranked diagnoses (e.g. ADHD) difficult. Therefore, the options to diagnose multiple disorders were extended, and diagnostic criteria were more specific and sensitive in later versions of the manual (Beuter & Malik, 2002). Now, the best next step would be to no longer rely on a categorical approach (i.e. the disorder is either present or absent), but rather to adopt a dimensional model where deficits can be conceptualized as falling somewhere along a continuum that ranges from normal to pathological. An important pioneer in this field is the National Institute of Mental Health (NIMH), which decided to no longer adhere to the current classification system, and to apply an experimental approach to the classification of mental disorders. The NIMH recently launched the Research Domain Criteria project (RDoC) to implement this strategy that incorporates not only behavioral symptoms, but also measures from neurocognitive, neurobiological and genetic research (http:// www.nimh.nih.gov/research-priorities/rdoc/index.shtml). The inclusion of multiple domains provides a broader view on developmental disorders in general, and a framework that ultimately brings the approach to disorders such as ASD and criteria long before it is time to start thinking about a DSM-6. Genetic research A bottom-up classification also allows for the inclusion of quantitative information on comorbid symptoms into genetic studies. Since the LCA also permit the inclusion of milder but still impairing (‘below the clinical cut-off’) behavior, it may improve correspondence between phenotypic variance and susceptibility genes. For example, Acosta et al. (2008) performed LCA on symptom measures for ADHD, oppositional defiant disorder (ODD), conduct disorder (CD), anxiety and depression to improve correspondence between phenotypic variance and susceptibility genes. Their findings indicated that a dimensional approach that includes milder but still impairing phenotypes (i.e. below the clinical cut-off) may even be essential for analyses in large molecular genetic studies, relevant for future clinical classifications, and may help resolve the contradictory results presented in current molecular genetic studies. Studies conducted by Li & Lee (2010; 2012) additionally illustrated that since LCA minimize the within-group heterogeneity seen in traditional diagnostic categories, they improve the grip on complex phenotypes in genetic studies. Unfortunately, the homogeneous groups within our current samples were too small to analyze susceptibility genes, but we are currently in the process of merging genetic, cognitive, and behavioral data from multiple research projects. Biomarkers in clinical practice ADHD closer to the development of more sophisticated treatment. Although these Thus far, neuroimaging studies mainly focused on cortical regions and their neurocognitive, neurobiological and genetic domains have not made it into the connections, and demonstrated global cortical maturation delay based on DSM-5, it is acknowledged that it may not be too early to use neurobiology as a reduced cortical thickness and reduced grey and white matter volumes in ADHD central tool to rethink the current approach to mental disorders (Hyman, 2007; (see for review de la Fuente et al., 2013). Future research may wish to put more 2010; Miller, 2010). That is, ongoing research could detach science from the effort on understanding the roles of the subcortical structures such as the basal 204 205 Chapter 8 ganglia, and their structural and functional pathways in ADHD, which may also individual’s experience of the learning environment would provide insight into significantly contribute to the pathophysiology of ADHD. A systematic review and how individual differences in brain development interact with formal education. meta-analysis of diffusion-weighted MRI studies (van Ewijk, Heslenfeld, Zwiers, Second, how the education is provided, for example with regard to the amount of Buitelaar & Oosterlaan, 2012) indeed described positive associations between on rehearsal offered, can be altered according to these individual differences. Ideally, the one hand ADHD-behavior or cognitive deficits typical for ADHD (e.g. attentional the educational system should be able to monitor and adapt to the learner’s focus, interference control, verbal fluency), and on the other hand white matter current repertoire of skills and knowledge. A promising approach involves the abnormalities in amongst others the inferior and superior longitudinal fasciculus, development of technology-enhanced learning applications that are capable and areas within the basal ganglia. of adapting to individual needs (for review see Butterworth & Kovas, 2013). Based on the qualitative differences between homogeneous ASD-ADHD Improving educational quality by meeting the needs of children with ASD and classes in visual-spatial task performances described in the chapters 3 and 5, ADHD is likely to positively affect their general wellbeing, since these children will studying brain activation during such tasks may be a promising future research experience less frustration and more success. Current research did not contain avenue. By comparing children with elevated scores on either ASD or ADHD, suitable measures to examine whether the non-symptomatic ends of the ASD children with elevated scores on both, and control children regarding their brain and ADHD traits were associated with positive constructs such as wellbeing and activation patterns during such visual-spatial tasks, (mal)adaptive structural and quality of life; this therefore remains an interesting question for future research. functional pathways unique or shared in ASD and ADHD may be detected. Of note, given that a task-transcending cognitive impairment (i.e. overall slower and Clinical implications more inaccurate performance) such as described in chapter 6 was an aspecific Given the high comorbidity rates between ASD and ADHD (an estimated but strong predictor of psychiatric dysfunctioning in general, future studies might 30% to 80% ADHD in ASD, and about 20% to 50% ASD in ADHD; for review also wish to examine how generic cognitive profiles map into the efficiency and see Rommelse et al., 2010), and given the homogeneous classes with mostly modularity of functional neural networks. Ultimately, the detection of biomarkers similar levels of both disorders detected across the general population (chapter for ASD and/or ADHD can be of clinical utility in children with an unclassifiable 5), findings suggest that children who are referred to a clinical centre for either behavioral profile, i.e. behavioral problems that are milder than full blown ASD or ADHD should be examined for symptoms of the other disorder as well. diagnoses, but still impairing. Such biomarkers can for instance provide extra As the clinical presentation of both disorders is strongly influenced by age, this information with regard to expected response to treatment. examination should occur on a regular basis (St. Pourcain et al., 2011). Well in line with this recommendation is that children with solely ADHD symptoms Life outside clinical practice – education and learning (i.e. the ADHD-only class described in chapter 3) were slightly younger than the Butterworth & Kovas (2013) described two ‘grand challenges’ that have to be met children who presented with both ASD and ADHD symptoms. These ADHD-only to understand disorders such as ASD and ADHD, and to improve education for children may have a childhood-limited form of ADHD, a hypothesis that calls children with ASD and ADHD and formal education in general. First, unraveling for regular examinations. As children with ‘pure’ ASD are rarely seen in clinical how cognitive processes, their neural basis, and genetic etiology influence the populations (despite the fact that we included children in our study on the basis 206 207 Chapter 8 of the presence of ASD), clinicians will most of the time see patients with ASD and comorbid ADHD. Therefore, clinicians should keep in mind that ADHD-specific difficulties may make treatment interventions harder to grasp for ASD-affected children. For instance, social skills training may be more difficult due to comorbid ADHD problems such as inattention. Literature reports greater impairment in adaptive functioning and a poorer health-related quality of life for children with ASD and clinically significant ADHD symptoms in comparison with children with ASD and fewer ADHD symptoms (Sikora et al., 2012). Therefore, clinicians may wish to adjust treatment plans to ensure comprehensive and effective treatment for both ASD and associated ADHD. In addition, the two subtypes of children with both ASD and ADHD behavior described in chapter 3(i.e. with one or the other more profound) differed not only with regard to comorbid internalizing and externalizing symptoms, but also qualitatively with regard to their visual-spatial skills and emotion recognition abilities. Therefore, these children may benefit from different clinical approaches. Based on the absence of an association between behavioral symptoms and inhibitory control (chapter 7), consideration should be given to the cognitive measures used in the assessment of medical treatment outcome in ASD and ADHD. Perhaps predicting or monitoring treatment response using cognitive tests may not always be of clinical utility. Furthermore, since the cognitive profiles in theclinic and population-based sample largely overlapped, and since these profiles were defined by similar strengths and difficulties across cognitive tasks, standard cognitive batteries may mainly tap into the shared underlying variance relevant to psychiatric dysfunctioning in general. Even though this is quite informative in its own right, as this shared cognitive variance may link to etiological factors relevant to dysfunctioning in general (Caspi et al., 2013; CrossDisorder Group of the Psychiatric Genomics, 2013), it does not bring us any closer to understanding the heterogeneity in ASD/ADHD affected populations, let alone in developing more tailored diagnostic and treatment approaches. 208 209 Summary in Dutch 210 Wat zijn Autismespectrumstoornis en Aandachtstekortstoornis met Hyperactiviteit? Autismespectrumstoornis (ASS) en Aandachtstekortstoornis met Hyperactiviteit (ADHD) zijn psychiatrische ontwikkelingsstoornissen. ASS wordt gekenmerkt door problemen in de sociale communicatie en repetitief en/of stereotiep gedrag en interesses (American Psychiatric Association, 2013). In het contact met een kind met ASS is er bijvoorbeeld weinig/geen sprake van wederkerigheid, en door onvoldoende begrip van humor en abstract taalgebruik is het voor kinderen met ASS moeilijker om een gesprek te voeren. Het repetitief en stereotiep gedrag blijkt bijvoorbeeld uit geobsedeerd bezig zijn met bepaalde voorwerpen of onderwerpen, vrijetijdsbestedingen en stereotiepe bewegingen. Voor ADHD zijn aandachtsproblemen en/of hyperactiviteit en impulsiviteit het meest kenmerkend (American Psychiatric Association, 2013). Aandachtsproblemen bij kinderen met ADHD zijn bijvoorbeeld waarneembaar als verhoogde afleidbaarheid van schoolwerk of dagdromen. Hyperactief en impulsief gedrag blijken bijvoorbeeld uit voortdurend in de weer zijn en anderen onderbreken tijdens het praten. ASS en ADHD zijn gedragsdiagnoses; ze worden vastgesteld op basis van diagnostische interviews en observaties door experts in de kinderen jeugdpsychiatrie. Leidraad voor dit diagnostisch onderzoek is de Diagnostic and Statistical Manual of Mental Disorders (DSM-5) (American Psychiatric Association, 2013). De DSM-5 is een classificatiesysteem waarin de precieze gedragskenmerken van onder andere ASS en ADHD staan omschreven. In dit classificatiesysteem zijn criteria vastgelegd aangaande de hoeveelheid symptomen, de mate van ernst van deze symptomen en het eerste voordoen van deze symptomen die voorwaarde zijn om een diagnose te mogen stellen. Op basis van deze criteria wordt bij ongeveer 1% van de kinderen ASS gediagnosticeerd en bij ongeveer 5% van de kinderen ADHD (Baird et al., 2006; Polanczyk et al., 2007). Hiermee zijn ASS en ADHD een van de meest voorkomende psychiatrische ontwikkelingsstoornissen wereldwijd. 213 Chapter 9 De waarde van de DSM is dat experts overal ter wereld over exact hetzelfde en unieke oorzakelijke factoren van beide stoornissen. In de recent verschenen praten en naar exact hetzelfde onderzoek doen wanneer zij het ASS of ADHD DSM-5 mogen beide diagnoses wel samen gesteld worden, hetgeen een impuls noemen. Vrijwel alle kennis die we nu over ASS en ADHD hebben is gebaseerd op kan geven aan wetenschappelijk onderzoek naar de oorzakelijke factoren voor volgens de DSM gedefinieerde groepen kinderen. Dit classificatiesysteem heeft ASS en ADHD. echter ook een belangrijk nadeel. De DSM is gebaseerd op consensus tussen experts aangaande categorieën en niet op de meest actuele wetenschappelijke nemen van een breder perspectief door gebruik te maken van dimensionele maten. inzichten. Het is onwaarschijnlijk dat deze DSM-categorieën goed afgrensbaar Bij een dimensionele benadering wordt geen afkappunt gehanteerd, maar wordt zijn op basis van oorzaken en prognose van bepaalde symptomen. Het afkappunt het gehele continuüm van gedragskenmerken meegenomen. Kinderen worden voor wanneer er wel/niet sprake is van een diagnose ASS of ADHD is arbitrair gekarakteriseerd op basis van de mate waarin zij verschillende symptomen zolang hiervoor geen objectieve criteria zijn. laten zien. Hierbij wordt geen onderscheid gemaakt tussen types symptomen, Dit probleem speelt op meerdere fronten. Ten eerste is het afkappunt tussen bijvoorbeeld of symptomen passen bij ASS of ADHD. Bij een dimensionele bijvoorbeeld ‘normale’ repetitieve gedragingen en ‘normale’ aandachtsproblemen beoordeling van gedragskenmerken worden alle kinderen, ongeacht de aan-of versus klinisch repetitief gedrag en klinische aandachtsproblemen arbitrair. ASS afwezigheid van DSM-classificaties, op meerdere dimensies beoordeeld. Een goede methodiek voor het detecteren van deze grondslagen is het en ADHD-gedragskenmerken zijn continu verdeeld in de algemene populatie. Dit betekent dat ook kinderen die geen DSM-classificatie hebben variëren in Neuropsychologie bij ASS en ADHD de aan-en afwezigheid van gedragskenmerken die passen bij ASS en ADHD. Neuropsychologie bestudeert de relatie tussen gedrag en breinfunctioneren, Door gebruik te maken van DSM-categorieën worden eventuele problemen bij hiermee kinderen die onder het afkappunt voor DSM-classificaties zitten miskend. Ten Neuropsychologen bestuderen de werking van de hersenen op gedragsniveau. tweede is niet iedereen die een ASS-diagnose heeft hetzelfde, en iedereen die Dat betekent dat neuropsychologen proberen te ontrafelen hoe onze hersenen een ADHD-diagnose heeft evenmin. Groepsvergelijkingen op basis van DSM- ons in staat stellen om ons gedrag te sturen. Veel neuropsychologisch onderzoek indelingen zijn hierdoor gebaseerd op heterogene groepen (zie bijvoorbeeld Lord richt zich hierbij op zogenaamde cognitieve functies zoals aandacht, geheugen & Jones, 2012; Miller, 2010). Dit probleem is mede ontstaan doordat de DSM- en cognitieve flexibiliteit. Deze neurocognitieve taken zijn bijvoorbeeld geschikt classificaties ASS en ADHD tot voor kort formeel niet samen gesteld mochten voor het maken van sterkte-zwakte profielen van kinderen. Kinderen met ASS en worden. Veel kinderen met een diagnose ASS laten echter ook symptomen zien ADHD-diagnoses zijn bekend met cognitieve problemen, en neuropsychologen die kenmerkend zijn voor ADHD, en vice versa laten ook relatief veel kinderen proberen verbanden te leggen tussen de gedragskenmerken enerzijds en de met een ADHD-diagnose symptomen zien die passen bij ASS. In de klinische cognitieve problemen anderzijds. Een cognitief probleem dat kenmerkend praktijk lopen de schattingen voor deze comorbiditeit uiteen; ongeveer 30% tot is voor ASS is bijvoorbeeld een zwakke centrale coherentie, wat betekent dat 80% van de kinderen met ASS heeft ook ADHD, en ongeveer 20% tot 50% van de kinderen met ASS gemiddeld genomen niet gehelen maar details waarnemen kinderen met ADHD heeft ook ASS (voor review zie Rommelse et al., 2011). Door (zie bijvoorbeeld Booth & Happé, 2010). Een cognitief probleem dat kenmerkend de eerdere DSM-restricties zijn tot dusverre weinig studies verricht naar gedeelde is voor ADHD is bijvoorbeeld een zwakke inhibitie, wat betekent dat kinderen met 214 is neuropsychologie een van de hersenwetenschappen. 215 Chapter 9 ADHD gemiddeld genomen meer moeite hebben met het remmen van gedrag (zie zijn in meer of mindere mate identiek in de patronen die zij vertonen. Zij laten bijvoorbeeld Crosbie et al., 2013). Naast cognitieve problemen die meer typisch bijvoorbeeld veel ADHD-gedragskenmerken gecombineerd met weinig ASS- zijn voor ASS of ADHD zijn er meerdere cognitieve problemen die bij zowel ASS gedragskenmerken zien, of zij laten bijvoorbeeld zien heel vlot, maar weinig als ADHD gerapporteerd worden (zie voor review Rommelse et al., 2011). accuraat te kunnen werken. Op basis van deze patronen kunnen kinderen Zowel het frequent samen voorkomen van ASS en ADHD- ingedeeld worden in subgroepen (classes) die meer homogeen zijn. Door meer gedragskenmerken en -diagnoses, als de overeenkomsten in cognitieve problemen homogene classes te vergelijken kunnen de grondslagen van ASS en ADHD- indiceert dat er mogelijk gelijke processen aan beide ontwikkelingsstoornissen gedragskenmerken of cognitieve profielen beter gedetecteerd worden. ten grondslag liggen. Deze unieke en gedeelde cognitieve factoren onderliggend aan ASS en ADHD-gedragskenmerken kunnen het veld wijzen in de richting van Wat zijn de belangrijkste bevindingen? meer objectieve diagnostiek. Net als voor gedragskenmerken geldt ook voor In hoofdstuk 2 bekijken we het continuüm van ASS en ADHD-symptomen in de cognitieve problemen dat deze continu verdeeld zijn in de algemene populatie. algemene populatie. We onderzoeken of de beide uiteinden van het continuüm Dit betekent dat ook kinderen die geen DSM-classificatie hebben variëren in hun (het onaangedane uiteinde zonder ASS en ADHD-gedragskenmerken en cognitieve sterke en zwakke kanten (zie bijvoorbeeld Fair et al., 2012). Door ook het aangedane uiteinde met ASS en ADHD-symptomen) meer kwantitatief cognitieve profielen dimensioneel te beschouwen, dus het hele continuüm mee dan kwalitatief van elkaar verschillen. De resultaten laten zien dat zowel de te nemen en geen afkappunt te hanteren, wordt ook deze cognitieve variatie in de hoeveelheid comorbide gedragsproblemen als (in mindere mate) het cognitief algemene populatie erkend. presteren gunstiger is met minder ASS en ADHD-symptomen; zowel qua gedrag als qua cognitie worden geen nadelen gevonden voor het vrij zijn van ASS en Het onderzoek in dit proefschrift ADHD-symptomen. Dit proefschrift heeft ten doel de cognitieve profielen die horen bij de symptomen van ASS en ADHD te ontrafelen. Welke cognitieve sterktes en zwaktes horen 3, 4, 5 en 6 de heterogeniteit op een van beide domeinen beperkt. Hiervoor wordt specifiek bij ASS-symptomen, welke horen specifiek bij ADHD-symptomen en gebruik gemaakt van gedragsmatig of cognitief meer homogene latente classes. welke passen bij allebei? De bijbehorende cognitieve profielen bieden mogelijk In hoofdstuk 3 wordt de hypothese getoetst of ADHD een minder ernstige variant meer inzicht in de gedeelde en unieke grondslagen van ASS en ADHD. binnen het ASS-spectrum is. Als dat klopt, dan zou er wel zoiets kunnen bestaan Het detecteren van deze grondslagen wordt in dit proefschrift nagestreefd als enkelvoudige ADHD, maar zal ASS altijd gepaard gaan met comorbide ADHD- door gebruik te maken van dimensionele data, afkomstig van kinderen (5 -17 symptomen. In de LCA wordt inderdaad wel een class met enkelvoudige ADHD jaar) uit zowel de algemene populatie als een klinische ASS-populatie. Deze en geen class met enkelvoudige ASS gevonden. Er wordt bovendien veel overlap dimensionele maten kunnen worden ingevoerd in zogenaamde latente klasse in cognitieve problemen passend bij ASS-symptomen en passend bij ADHD- analyses (latent class analyses: LCA) (McCutcheon, 1987). LCA is een data- symptomen gevonden, bijvoorbeeld voor motorische vaardigheid, aandacht en analysemethode waarbij ervan uit wordt gegaan dat er patronen bestaan in de emotieherkenning. Grofweg hangt een toename in ASS en ADHD-symptomen data, bijvoorbeeld in gedragskenmerken of cognitieve symptomen. Kinderen samen met een toename in cognitieve problemen. Echter, het visuo-spatieel 216 Aangaande de relatie tussen gedrag en cognitie wordt in de hoofdstukken 217 Chapter 9 functioneren blijkt verschillend voor kinderen met voornamelijk ASS-symptomen ASS en ADHD stelt het klinisch nut van uitgebreide neurocognitieve taakbatterijen en kinderen met voornamelijk ADHD-symptomen. bij de diagnostiek naar ASS en ADHD ter discussie. Een zwakke motorische timing wordt in verband gebracht met primaire In hoofdstuk 7 komt tenslotte een dubbelblind placebo-gecontroleerde ADHD-symptomen zoals moeite met het remmen van gedrag en het richten en klinische studie onder kinderen met zowel een ASS-diagnose als ADHD- vasthouden van aandacht. De studie beschreven in hoofdstuk 4 laat zien dat symptomen aan bod. In deze studie wordt het effect van de noradrenaline- de motorische timing ernstiger is aangedaan bij kinderen die niet alleen ADHD- heropnameremmer atomoxetine op de prestaties op twee inhibitietaken symptomen, maar ook ASS-symptomen presenteren. Dit resultaat sluit aan bij de bestudeerd. De resultaten laten zien dat een afname in ADHD-symptomen niet hypothese dat ADHD een minder ernstige variant binnen het ASS-spectrum is. samenhangt met een verbetering in inhibitie. Deze bevinding kan betekenen In hoofdstuk 5 worden, evenals in hoofdstuk 2, alleen kinderen uit dat er verschillende oorzaken aan ADHD-symptomen en inhibitieproblemen bij de algemene populatie bestudeerd. In deze studie worden gedragsmaten kinderen met ASS en ADHD ten grondslag liggen. Er kan geconcludeerd worden gebruikt die sensitief zijn voor de variatie in ASS-gedragskenmerken en ADHD- dat dit type cognitieve taken niet geschikt is voor het monitoren van vooruitgang gedragskenmerken in de algemene populatie. Uit de resultaten komt naar voren in ADHD-gedragskenmerken bij kinderen met zowel ASS als ADHD-problematiek. dat de gedragsprofielen en cognitieve profielen binnen de algemene populatie grotendeels overeenkomen met de resultaten binnen de klinische populatie. Concluderend Hieruit kan geconcludeerd worden dat een deel van de heterogeniteit die • Een dimensionele benadering is geschikt voor het detecteren van associaties kenmerkend is voor ontwikkelingsstoornissen als ASS en ADHD, ook gevonden kan worden binnen de algemene populatie. tussen gedragskenmerken en cognitief functioneren. • In hoofdstuk 6 wordt een innovatieve aanpak gehanteerd, waarbij homogene classes gebaseerd worden op cognitieve maten in plaats van populatie is te herleiden tot variatie in de algemene populatie. • gedragsmaten. Uit deze studie komt naar voren dat de cognitieve profielen van kinderen met en zonder ASS en ADHD-symptomen grotendeels vergelijkbaar zijn. Een deel van de gedragsmatige en cognitieve variatie binnen de klinische ASS-gedragskenmerken gaan gepaard met ADHD-gedragskenmerken, terwijl ADHD-gedragskenmerken ook frequent enkelvoudig voorkomen. • Een overwegend ASS-profiel en een overwegend ADHD-profiel verschillen Er worden geen cognitieve sterktes en zwaktes gevonden op specifieke domeinen, qua cognitieve prestaties voornamelijk in kwantitatieve zin van elkaar, ten maar profielen die te onderscheiden zijn op basis van specifieke combinaties van nadele van het ASS-profiel. snelheid en accuratesse over de verschillende domeinen. Classes die minder • Enkel het visueel-spatieel functioneren verschilt kwalitatief voor kinderen met accuraat en langzamer presteren laten meer ASS en ADHD-symptomen zien, en overwegend ASS-gedragskenmerken en kinderen met overwegend ADHD- classes die accurater en vlotter presteren laten minder ASS en ADHD-symptomen gedragskenmerken, met meer oog voor detail bij een overwegend ASS- zien. Er is sprake van vergelijkbare heterogeniteit binnen de klinische populatie profiel. en algemene populatie, met kwantitatieve meer dan kwalitatieve verschillen over het continuüm. Het vinden van meer aspecifieke cognitieve voorspellers voor 218 • Cognitieve profielen wijzen op een generiek minder accuraat en minder vlot presteren bij meer ASS en ADHD-gedragskenmerken. 219 Chapter 9 • Het maken van cognitieve sterkte-zwakte profielen en het monitoren van gedragsmatige vooruitging middels cognitieve taken vereist een duidelijke samenhang tussen gedrag en cognitie. Welke richting geven deze bevindingen voor de toekomst? Uit dit proefschrift zijn twee hoofdbevindingen te destilleren die richting kunnen geven aan toekomstig wetenschappelijk onderzoek en de toekomstige behandeling van kinderen met ASS en ADHD. Ten eerste is gebleken dat door gebruik te maken van dimensionele maten voor ASS en ADHD, associaties tussen gedrag en cognitie naar voren komen die met een DSM-groepsindeling mogelijk niet gedetecteerd zouden worden. Deze kennis pleit voor een breder perspectief in toekomstig wetenschappelijk onderzoek, zoals ook geïnitieerd door het National Institute of Mental Health (NIMH) in het Research Domain Criteria project (RDoC). Hierbij wordt niet uitgegaan van DSM-classificaties op basis van gedragskenmerken, maar van meer objectiveerbare domeinen zoals het cognitief functioneren, het breinfunctioneren en de genetica. Een toekomstig classificatiesysteem zou dan niet langer een categoriale indeling hebben, maar gebaseerd worden op multidimensionele uitkomstmaten met meer constructvaliditeit. Over de keuze van geschikte uitkomstmaten moet consensus bestaan, en hierbij moet rekening gehouden worden met de context en ontwikkeling van de stoornissen. Ten tweede laten de resultaten zien dat de cognitieve en gedragsmatige heterogeniteit binnen klinische populaties met ASS en ADHD-symptomen deels geworteld zijn in vergelijkbare heterogeniteit binnen de algemene populatie. Eventuele verschillen tussen de niet-symptomatische en symptomatische uiteinden van de ASS en ADHD-continua blijken veelal kwantitatief van aard te zijn. Dit inzicht kan aanleiding zijn tot het bestuderen van mogelijke beschermende factoren voor het ontwikkelen van ASS en ADHD-gedragskenmerken binnen de algemene populatie. Wanneer beschermende factoren gevonden worden biedt dit mogelijke aanknopingspunten voor preventieve interventies. 220 221 References 222 A Acosta MT, Castellanos FX, Bolton KL, Balog JZ, Eagen P, Nee L, et al. (2008). Latent class subtyping of attention-deficit/ hyperactivity disorder and comorbid conditions. Journal of the American Academy of Child & Adolescent Psychiatry, 47(7), 797-807. Adamo N, Huo L, Adelsberg S, Petkova E, Castellanos FX & Di Martino A (2014). Response time intra-subject variability: commonalities between children with autism spectrum disorders and children with ADHD. European Child & Adolescent Psychiatry, 23(2), 69-79. Alderson RM, Rapport MD & Kofler MJ (2007). Attention-deficit/hyperactivity disorder and behavioral inhibition: a meta-analytic review of the stopsignal paradigm. Journal of Abnormal Child Psychology, 35(5), 745-758. Allman MJ, DeLeon IG & Wearden JH (2011). Psychophysical assessment of timing in individuals with autism. American Journal on Intellectual and Developmental Disabilities, 116(2), 165-178. Althoff RR, Copeland WE, Stanger C, Derks EM, Todd RD, Neuman RJ, et al. (2006). The latent class structure of ADHD is stable across informants. Twin Research and Human Genetics, 9(4), 507-522. American Psychiatric Association (2013). Diagnostic and Statistical Manual of Mental Disorders (5thEd.). Arlington, VA: American Psychiatric Publishing. American Psychiatric Association (2000). Diagnostic and Statistical Manual of Mental Disorders, Text Revision, (4th Ed.). Washington, DC: American Psychiatric Association. Ames CS & White SJ (2011). Are ADHD traits dissociable from the autistic profile? Links between cognition and behaviour. Journal of Autism and Developmental Disorders, 41(3), 357-363. 225 Chapter 10 Andreou P, Neale BM, Chen W, Christiansen H, Gabriels I, Heise A, et al. (2007). Reaction time performance in ADHD: improvement under fast-incentive condition and familial effects. Psychological Medicine, 37(12), 17031715. A common variant of the latrophilin 3 gene, LPHN3, confers susceptibility to ADHD and predicts effectiveness of stimulant medication. Molecular Psychiatry, 15(11), 1053-1066. (2006). Prevalence of disorders of the autism spectrum in a population RK (2011). The SWAN Captures Variance at the Negative and Positive Ends of the ADHD Symptom Dimension. Journal Attention Disorders, 17(2), 152-162. Arnold LE, Aman MG, Cook AM, Witwer AN, Hall KL, Thompson S & Ramadan Y (2006). Atomoxetine for hyperactivity in autism spectrum disorders: placebo-controlled crossover pilot trial. Journal of the American Academy of Child & Adolescent Psychiatry, 45(10), 1196-1205. relevance for genetic research in attention-deficit/hyperactivity disorder. Biological Psychiatry, 57(11), 1285-1292. tic disorders & ADHD: conceptual and methodological considerations. Methylphenidate enhances prepulse inhibition during processing attention-deficit/hyperactivity disorder. Psychophysiology, 47(5), 838-845. Auyeung B, Baron-Cohen S, Wheelwright S & Allison C (2008) The autism spectrum quotient: Children’s version (AQ-child). Journal of Autism and Developmental Disorders, 38 (7), 1230-1240. Bari A, Eagle DM, Mar AC, Robinson ES & Robbins TW (2009). Dissociable effects of noradrenaline, dopamine, and serotonin uptake blockade on stop task performance in rats. Psychopharmacology, 205(2), 273-283. Barkley RA (2011). Barkley Deficits in Executive Functioning Scale (BDEFS) New York, NY: Guilford Press. Barkley RA&Fischer M (2011) Predicting impairment in major life activities reported executive function (EF) deficits versus EF tests. Developmental Neuropsychology, 36(2), 137-161. Barkley RA (1997). Behavioural inhibition, sustained attention, and executive Ashare RL, Hawk LW,Shiels K, Rhodes JD, Pelham WE & Waxmonsky JG (2010). in Banaschewski T, Neale BM, Rothenberger A & Roessner V (2007)Comorbidity of and occupational functioning in hyperactive children as adults: self- Aron AR & Poldrack RA (2005). The cognitive neuroscience of response inhibition: stimuli Project (SNAP). The Lancet, 368 (9531), 210-215. European Child & Adolescent Psychiatry, 16(9), 5-14. Arnett AB, Pennington BF, Friend A, Willcutt EG, Byrne B, Samuelsson S & Olson task-relevant Baird G, Simonoff E, Pickles A, Chandler S, Loucas T, Meldrum D & Charman T cohort of children in South Thames: the Special Needs and Autism Arcos-Burgos M, Jain M, Acosta MT, Shively S, Stanescu H, Wallis D, et al. (2010). of B functions: constructing a unifying theory of AD/HD. Psychological Bulletin, 121(1), 65-94. Barkley RA, Grodzinsky G & DuPaul GJ (1992). Frontal lobe functions in attention deficit disorder with and without hyperactivity: a review and research report. Journal of Abnormal Child Psychology, 20(2), 163-188. Barnett JH, Heron J, Goldman D, Jones PB & Xu K (2009). Effects of catecholO-methyltransferase on normal variation in the cognitive function of children. American Journal of Psychiatry, 166(8), 909-916. 226 227 Chapter 10 Baron-Cohen S, Wheelwright S, Skinner R, Martin J & Clubley E (2001). The autism-spectrum quotient (AQ): evidence from Asperger syndrome/high- symptoms in the general population at 18 months. Journal of Autism and Developmental Disorders, 43(1), 45-56. functioning autism, males and females, scientists and mathematicians. Journal of Autism and Developmental Disorders, 31(1), 5-17. Beutler LE & Malik ML (2002). Rethinking the DSM: A Psychological Perspective. American Psychological Asoociation, Washington DC. Bauermeister JJ, Shrout PE, Ramirez R, Bravo M, Alegría M, Martínez-Taboas A, et al. (2007). ADHD correlates, comorbidity, and impairment in community and treated samples of children and adolescents. Journal of Abnormal Blondeau C & Dellu-Hagedorn F (2007). Dimensional analysis of ADHD subtypes in rats. Biological Psychiatry, 61(12), 1340-1350. Child Psychology, 35(6), 883-898. Boonstra AM, Kooij JJ, Oosterlaan J, Sergeant JA & Buitelaar JK (2005). Does Belmonte MK, Allen G, Beckel-Mitchener A, Boulanger LM, Carper RA & Webb SJ methylphenidate improve inhibition and other cognitive abilities in (2004). Autism and abnormal development of brain connectivity. Journal adults with childhood-onset ADHD? Journal of Clinical and Experimental of Neuroscience, 24(42), 9228-9231. Neuropsychology, 27(3), 278-298. Benjamini Y & Hochberg Y (1995). Controlling the false discovery rate: a practical Booth R & Happé F (2010). “Hunting with a knife and . . . fork”: examining and powerful approach to multiple testing. Journal of the Royal Statistical central coherence in autism, attention deficit/hyperactivity disorder, and Society: Series B, 57, 289-300. typical development with a linguistic task. Journal of Experimental Child Psychology, 107(4), 377-393. Benvenuto A, Battan B, Porfirio MC & Curatolo P (2012). Pharmacotherapy of autism spectrum disorders. Brain &Development, 35(2),119-127. Booth R, Charlton R, Hughes C & Happé F (2003). Disentangling weak coherence and executive dysfunction: planning drawing in autism and attention- Bekker EM, Overtoom CC, Kenemans JL, Kooij J, De Noord I, Buitelaar JK & Verbaten MN (2005). Stopping and changing in adults with ADHD. deficit/hyperactivity disorder. Philosophical Transactions of the Royal Society B: Biological Sciences, 358(1430), 387-392. Psychological Medicine, 35, 807-816. Bunge SA, Hazeltine E, Scanlon MD, Rosen AC & Gabrieli JD (2002). Dissociable Benjamini Y & Hochberg Y (1995). Controlling the false discovery rate: a practical and powerful approach to multiple testing. Journal of the Royal Statistical contributions of prefrontal and parietal cortices to response selection. Neuroimage, 17(3), 1562-1571. Society: Series B, 57, 289-300. Bush G, Valera EM & Seidman LJ (2005). Functional neuroimaging of attentionBernfeld J (2012). ADHD and factor analysis: are there really three distinct subtypes of ADHD? Applied Neuropsychology Child, 1(2), 100-104. Beuker KT, Schjolberg S, Kveim Lie K, Donders R, Lappenschaar M, Swinkels SH & Buitelaar JK (2013). The structure of autism spectrum disorder 228 deficit/hyperactivity disorder: a review and suggested future directions. Biological Psychiatry, 57(11), 1273-1284. Butterworth B & Kovas Y (2013). Understanding neurocognitive developmental disorders can improve education for all. Science, 340 (6130), 300-305. 229 Chapter 10 Booth R & Happé F (2010). “Hunting with a knife and . . . fork”: examining Bush G, Valera EM & Seidman LJ (2005). Functional neuroimaging of attention- central coherence in autism, attention deficit/hyperactivity disorder, and deficit/hyperactivity disorder: a review and suggested future directions. typical development with a linguistic task. Journal of Experimental Child Biological Psychiatry, 57(11), 1273-1284. Psychology, 107, 377-393. C Booth R, Charlton R, Hughes C & Happé F (2003). Disentangling weak coherence and executive dysfunction: planning drawing in autism and attentiondeficit/hyperactivity disorder. Philosophical Transactions of the Royal Society B: Biological Sciences, 358 (1430), 387-392. Bradley RA & Srivastava, SS (1979). Correlation in polynomial regression. The American Statistician, 33(1), 11-14. Bralten J, Franke B, Waldman I, Rommelse N, Hartman C, Asherson P, et al.(2013). Candidate genetic pathways for attention-deficit/hyperactivity disorder (ADHD) show association to hyperactive/impulsive symptoms in children with ADHD. Journal of the American Academy of Child & Adolescent Psychiatry, 52(11), 1204-1212. Brieber S, Neufang S, Bruning N, Kamp-Becker I, Remschmidt H, HerpertzDahlmann B, Fink GR, et al. (2007). Structural brain abnormalities in adolescents with autism spectrum disorder and patients with attention deficit/hyperactivity disorder. Journal of Child Psychology and Psychiatry, 48(12), 1251-1258. Buhler E, Bachmann C, Goyert H, Heinzel-Gutenbrunner M & Kamp-Becker I (2011). Differential diagnosis of autism spectrum disorder and attention deficit hyperactivity disorder by means of inhibitory control and ’theory of mind’. Journal of Autism and Developmental Disorders, 41(12), 17181726. Bunge SA, Hazeltine E, Scanlon MD, Rosen AC & Gabrieli JD (2002). Dissociable contributions of prefrontal and parietal cortices to response selection. Cadesky EB, Mota VL & Schachar RJ (2000). Beyond words: how do children with ADHD and/or conduct problems process nonverbal information about affect? Journal of the American Academy of Child & Adolescent Psychiatry, 39(9), 1160-1167. Canitano R & Scandurra V (2011). Psychopharmacology in autism: an update. Progress in Neuropsychopharmacology and Biological Psychiatry, 35(1), 18-28. Caron C & Rutter M (1991). Comorbidity in child psychopathology: concepts, issues and research strategies. Journal of Child Psychology and Psychiatry, 32(7), 1063-1080. Casey BJ, Thomas KM, Welsh TF, Badgaiyan RD, Eccard CH, Jennings JR & Crone EA (2000). Dissociation of response conflict, attentional selection, and expectancy with functional magnetic resonance imaging. Proceedings of the National Academy of Sciences of the United States of America, 97(15), 8728-8733. Caspi A, Houts RM, Belsky BW, Goldman-Mellor SJ, Harrington H, Israel S, Meier MH, et al. (2013). The p-Factor: One General Psychopathology Factor in the Structure of Psychiatric Disorders? Clinical Psychological Science, doi: 10.1177/2167702613497473. Castellanos FX & Tannock R (2002). Neuroscience of attention-deficit/hyperactivity disorder: the search for endophenotypes. Nature Reviews Neuroscience, 3(8), 617-628. NeuroImage, 17, 1562-1571. 230 231 Chapter 10 Chamberlain SR, Robbins TW, Winder-Rhodes S, Muller U, Sahakian BJ, Conners CK, Sitarenios G, Parker JD & Epstein JN (1998b). Revision and Blackwell AD & Barnett JH (2011). Translational approaches to restandardization of the Conners Teacher Rating Scale (CTRS-R): factor frontostriatal dysfunction in attention-deficit/hyperactivity disorder using structure, reliability, and criterion validity. Journal of Abnormal Child a computerized neuropsychological battery. Biological Psychiatry, Psychology, 26(4), 279-291. 69(12), 1192-1203. Constantino JN (2011). The quantitative nature of autistic social impairment. Chamberlain SR, Del Campo N, Dowson J, Muller U, Clark L, Robbins TW & Pediatric Research, 69, 55-62. Sahakian BJ (2007). Atomoxetine improved response inhibition in adults with attention deficit/hyperactivity disorder. Biological Psychiatry, 62(9), Constantino JN, Abbacchi AM, Lavesser PD, Reed H, Givens L & Chiang L (2009). Developmental course of autistic social impairment in males. 977-984. Development and Psychopathology, 21(1), 127-138. Charnsil C (2011). Efficacy of Atomoxetine in Children With Severe Autistic Disorders and Symptoms of ADHD: An Open-Label Study. Journal of Attention Disorders,15(8), 684-689. Corbett BA, Constantine LJ, Hendren R, Rocke D & Ozonoff S (2009). Examining executive functioning in children with autism spectrum disorder, attention deficit hyperactivity disorder and typical development. Psychiatry Coghill DR, Hayward D, Rhodes SM, Grimmer C & Matthews K (2013) A longitudinal Research, 166(2-3), 210-222. examinationof neuropsychological and clinical functioning in boys with attention deficit hyperactivity disorder (ADHD): improvements in executive functioning do not explain clinical improvement. Psychological Cormier E (2008). Attention deficit/hyperactivity disorder: a review and update. Journal of Pediatric Nursing, 23(5), 345-357. Medicine, 44(5), 1087-1099. Cortese S, Castelnau P, Morcillo C, Roux S & Bonnet-Brilhault F (2012). Coghill DR, Seth S, Pedroso S, Usala T, Currie J & Gagliano A (2013). Effects of Psychostimulants for ADHD-like symptoms in individuals with autism Methylphenidate on Cognitive Functions in Children and Adolescents spectrum disorders. Expert Review of Neurotherapeutics, 12(4), 461-473. with Attention-Deficit/Hyperactivity Disorder: Evidence from a Systematic Review and a Meta-Analysis. Biological Psychiatry, doi: 10.1016/j. Crosbie J, Arnold P, Paterson A, Swanson J, Dupuis A, Li X, et al. (2013) Response inhibition and ADHD traits: Correlates and heritability in a community biopsych.2013.10.005. sample. Journal of Child Psychology and Psychiatry, 41(3), 497-507. Cohen J (1988). Statistical Power Analysis for the Behavioral Sciences, 2 Edition. nd Hillsdale, NJ: Erlbaum. Cross-Disorder Group of the Psychiatric Genomics Consortium (2013). Identification of risk loci with shared effects on five major psychiatric Conners CK, Sitarenios G, Parker JD & Epstein JN (1998a). The revised Conners’ disorders: a genome-wide analysis. The Lancet, 381 (9875), 1371-1379. Parent Rating Scale (CPRS-R): factor structure, reliability, and criterion validity. Journal of Abnormal Child Psychology, 26(4), 257-268. Cross-Disorder Group of the Psychiatric Genomics Consortium (2013). Genetic relationship between five psychiatric disorders estimated from genomewide SNPs. Nature Genetics, 45(9), 984-994. 232 233 Chapter 10 D Durston S (2003). A review of the biological bases of ADHD: what have we learned from imaging studies? Mental Retardation and Developmental Davis NO & Kollins SH (2012). Treatment for co-occurring attention deficit/ Disabilities Research Reviews, 9(3), 184-195. hyperactivity disorder and autism spectrum disorder. Neurotherapeutics, 9(3), 518-530. E Delucchi KL & Bostrom A (2004). Methods for analysis of skewed data distributions Economidou D, Theobald DE, Robbins TW, Everitt BJ & Dalley JW (2012). in psychiatric clinical studies: Working with many zero values. American Norepinephrine and Dopamine Modulate Impulsivity on the Five-Choice Journal of Psychiatry, 161(7), 1159-1168. Serial Reaction Time Task Through Opponent Actions in the Shell and Core Sub-Regions of the Nucleus Accumbens. Neuropsychopharmacology, 37(9), 2057-2066. Dennis M, Francis DJ, Cirino PT, Schachar R, Barnes MA & Fletcher JM (2009). Why IQ is not a covariate in cognitive studies of neurodevelopmental disorders. Journal of the International Neuropsychological Society, 15(3), Elia J, Arcos-Burgos M, Bolton KL, Ambrosini PJ, Berrettini W & Muenke M (2009). ADHD latent class clusters: DSM-IV subtypes and comorbidity. 331-343. Psychiatry Research, 170(2-3), 192-198. Di Martino A, Zuo XN, Kelly C, Grzadzinski R, Mennes M, Schvarcz A, et al. (2013). Shared and Distinct Intrinsic Functional Network Centrality in Autism Epstein JN, Conners CK, Hervey AS, Tonev ST, Arnold LE, Abikoff HB, et al. and Attention-Deficit/Hyperactivity Disorder. Biological Psychiatry, 74(8), (MTA Cooperative Study Group) (2006). Assessing medication effects 623-632. in the MTA study using neuropsychological outcomes. Journal of Child Psychology and Psychiatry, 47(5), 446-456. Doyle CA & McDougle CJ (2012). Pharmacotherapy to control behavioral symptoms in children with autism. Expert Opinion of van Ewijk H, Heslenfeld DJ, Zwiers MP, Buitelaar JK & Oosterlaan J (2012). Diffusion tensor imaging in attention deficit/hyperactivity disorder: A Pharmacotherapy13(11),1615-1629. systematic review and meta-analysis. Neuroscience & Biobehavioral Reviews, 36(4),1093–1106. Doyle AE (2006). Executive functions in attention-deficit/hyperactivity disorder. Journal of Clinical Psychiatry, 67(8), 21-26. F DuPaul GJ, Power TJ, Anastopoulos AD & Reid R (1998). ADHD Rating Scale-IV. Checklists, Norms, and Clinical Interpretations. New York: The Guilford Press. Durston S, Mulder M, Casey BJ, Ziermans T & van Engeland H (2006). Activation in ventral prefrontal cortex is sensitive to genetic vulnerability for attentiondeficit hyperactivity disorder. Biological Psychiatry, 60(10), 1062-1070. 234 Fair DA, Bathula D, Nikolas MA & Nigg JT (2012). Distinct neuropsychological subgroups in typically developing youth inform heterogeneity in children with ADHD. Proceeding of the National Academy of Sciences of the United States of America, 109(17) 6769-6774. Falter CM & Noreika V (in press). Time processing in developmental disorders: A comparative view. V Arstila & D Lloyd (Eds.) Subjective time: The 235 Chapter 10 philosophy, psychology, and neuroscience of temporality. Cambridge, Franke B, Faraone SV, Asherson P, Buitelaar J, Bau CH, Ramos-Quiroga JA, et MA: The MIT Press. al. (International Multicentre persistent ADHD CollaboraTion; IMPACT) (2012). The genetics of attention deficit/hyperactivity disorder in adults, Falter CM, Noreika V, Wearden JH & Bailey AJ (2012). More consistent, yet less a review. Molecular Psychiatry, 17(10), 960-987. sensitive: Interval timing in autism spectrum disorders. Quarterly Journal of Experimental Psychology, 65(11): 2093-2107. Franke B, Neale BM & Faraone SV (2009). Genome-wide association studies in ADHD. Human Genetics, 126(1), 13-50. Falter CM & Noreika V (2011). Interval timing deficits and abnormal cognitive development. Frontiers in Integrative Neuroscience, 5, 26-31. Frazier-Wood AC, Bralten J, Arias-Vasquez A, Luman M, Oosterlaan J, Sergeant J, et al. (2012). Neuropsychological intra-individual variability explains Faraone SV, Biederman J, Spencer T, Michelson D, Adler L, Reimherr F & unique genetic variance of ADHD and shows suggestive linkage to Seidman L (2005). Atomoxetine and stroop task performance in adult chromosomes 12, 13, and 17. American Journal of Medical Genetics attention-deficit/hyperactivity disorder. Journal of Child and Adolescent Part B: Neuropsychiatric Genetics, 159B, 131-140. Psychopharmacology, 15(4), 664-670. Freitag CM (2007). The genetics of autistic disorders and its clinical relevance: a Faraone SV, Perlis RH, Doyle AE, Smoller JW, Goralnick JJ, Holmgren MA & Sklar review of the literature. Molecular Psychiatry, 12(1), 2-22. P (2005). Molecular genetics of attention-deficit/hyperactivity disorder. Biological Psychiatry, 57(11), 1313-1323. de la Fuente A, Xia S, Branch C & Li X (2013). A review of attention-deficit/ hyperactivity disorder from the perspective of brain networks. Frontiers Fein D, Dixon P, Paul J & Levin H (2005). Brief report: pervasive developmental in Human Neuroscience, 15(7), 192-198. disorder can evolve into ADHD: case illustrations. Journal of Autism and Developmental Disorders, 35(4), 525-534. Fernández-Jaén A, Fernandez-Mayoralas DM, Calleja-Perez B, Munoz-Jareno N, Campos Diaz MD & Lopez-Arribas S (2013). Efficacy of Atomoxetine for the Treatment of ADHD Symptoms in Patients With Pervasive Developmental Disorders: A Prospective, Open-Label Study. Journal of Attentional Disorders, 17(6), 497-505. Fine JG, Semrud-Clikeman M, Butcher B & Walkowiak J (2008). Brief report: attention effect on a measure of social perception. Journal of Autism and Developmental Disorders, 38(9), 1797-1802. G Gadow KD, DeVincent CJ & Schneider J (2009). Comparative study of children with ADHD only, autism spectrum disorder + ADHD, and chronic multiple tic disorder + ADHD. Journal of Attentional Disorders, 12(5), 474-485. Gadow KD, DeVincent CJ & Drabick DA (2008). Oppositional defiant disorder as a clinical phenotype in children with autism spectrum disorder. Journal of Autism and Developmental Disorders, 38(7), 1302-1310. Gadow KD, DeVincent CJ & Pomeroy J (2006). ADHD symptom subtypes in children with pervasive developmental disorder. Journal of Autism and Developmental Disorders, 36(2), 271-283. 236 237 Chapter 10 Gargaro BA, Rinehart NJ, Bradshaw JL, Tonge BJ & Sheppard DM (2011). Goldstein S & Schwebach AJ (2004). The comorbidity of Pervasive Developmental Autism and ADHD: how far have we come in the comorbidity debate? Disorder and Attention Deficit Hyperactivity Disorder: results of a Neuroscience & Biobehavioral Reviews, 35(5), 1081-1088. retrospective chart review. Journal of Autism and Developmental Disorders, 34(3), 329-339. Gau SS & Shang CY (2010). Improvement of executive functions in boys with attention deficit hyperactivity disorder: an open-label follow- Gottesman II & Gould TD (2003). The endophenotype concept in psychiatry: up study with once-daily atomoxetine. International Journal of etymology and strategic intentions. American Journal of Psychiatry, Neuropsychopharmacology, 13(2), 243-256. 160(4), 636-645. Gepner B & Feron F (2009). Autism: a world changing too fast for a mis-wired brain? Neuroscience and Biobehavioral Reviews, 33(8), 1227-1242. Graf H, Abler B, Freudenmann R, Beschoner P, Schaeffeler E, Spitzer M, et al. (2011). Neural correlates of error monitoring modulated by atomoxetine in healthy volunteers. Biological Psychiatry, 69(9), 890-897. Geurts HM, van den Bergh SF & Ruzzano L (epub ahead of print). Prepotent Response Inhibition and Interference Control in Autism Spectrum Disorders: Two Meta-Analyses. Autism Research. Grinter EJ, Maybery MT, Van Beek PL, Pellicano E, Badcock JC & Badcock DR (2009). Global visual processing and self-rated autistic-like traits. Journal of Autism and Developmental Disorders, 39(9), 1278-1290. Geurts HM, Grasman RP, Verté S, Oosterlaan J, Roeyers H, van Kammen SM & Sergeant JA (2008). Intra-individual variability in ADHD, autism spectrum Grodzinsky GM & Diamond R (1992). Frontal lobe functioning in boys with disorders and Tourette’s syndrome. Neuropsychologia, 46(13), 3030- attention-deficit hyperactivity disorder. Developmental Neuropsychology, 3041. 8(4), 427-445. Geurts HM, Verté S, Oosterlaan J, Roeyers H & Sergeant JA (2004). How specific Groom MJ, Scerif G, Liddle PF, Batty MJ, Liddle EB, Roberts KL, et al. (2010). are executive functioning deficits in attention deficit hyperactivity Effects of motivation and medication on electrophysiological markers disorder and autism? Journal of Child Psychology and Psychiatry and of response inhibition in children with attention-deficit/hyperactivity Allied Disciplines, 45(4), 836-854. disorder. Biological Psychiatry, 67(7), 624-631. Ghanizadeh A (2012). Atomoxetine for Treating ADHD Symptoms in Autism: A Systematic Review. Journal of Attention Disorders, 17(8), 635-640. Groth-Marnat G (1997). Handbook of Psychological Assessment, 3rd Ed. New York: Wiley. Ghanizadeh A (2010). Factor analysis on ADHD and autism spectrum disorder Grzadzinski R, Di Martino A, Brady E, Mairena MA, O’Neale M, Petkova E, et DSM-IV-derived items shows lack of overlap. European Child and al.(2011). Examining autistic traits in children with ADHD: does the Adolescent Psychiatry, 19(10), 797-798. autism spectrum extend to ADHD? Journal of Autism and Developmental Disorders, 41(9), 1178-1191. 238 239 Chapter 10 Gualtieri CT & Johnson LG (2008). Medications do not necessarily normalize Happé F & Frith U (2006). The weak coherence account: detail-focused cognitive cognition in ADHD patients. Journal of Attention Disorders, 11(4), 459- style in autism spectrum disorders. Journal of Autism and Developmental 469. Disorders, 36(1), 5-25. Gueorguieva R & Krystal JH (2004). Move Over ANOVA: Progress in Analyzing Repeated-Measures Data and Its Reflection in Papers Published in the Happé F (1993). Autism: cognitive deficit or cognitive style? Trends in Cognitive Sciences, 3, 216-222. Archives of General Psychiatry. Archives of General Psychiatry, 61(3), 310-317. Harfterkamp M, van de Loo-Neus G, Minderaa RB, van der Gaag RJ, Escobar R, Schacht A, et al. (2012). A randomized double-blind study of atomoxetine H versus placebo for attention-deficit/hyperactivity disorder symptoms Hallmayer J, Cleveland S, Torres A, Phillips J, Cohen B, Torigoe T, et al. (2011). Genetic heritability and shared environmental factors among twin pairs with autism. Archives of General Psychiatry, 68(11), 1095-1102. outcome in adolescents/young adults with childhood ADHD: profiles of persisters, remitters and controls. Journal of Child Psychology and Psychiatry, 49(9), 958-966. Hamer RM & Simpson PM (2009). Last Observation Carried Forward Versus Mixed Models in the Analysis of Psychiatric Clinical Trials. American Journal of Psychiatry, 166, 639-641. Handen BL, Taylor J & Tumuluru R (2011). Psychopharmacological treatment of ADHD symptoms in children with autism spectrum disorder. International Journal of Adolescent Medicine and Health, 23(3), 167-173. Hanwella R, Senanayake M & de Silva V (2011). Comparative efficacy and acceptability of methylphenidate and atomoxetine in treatment of attention deficit hyperactivity disorder in children and adolescents: a meta-analysis. BMC Psychiatry, 11, 176. Academy of Child & Adolescent Psychiatry, 51(7), 733-741. Hay DA, Bennett KS, Levy F, Sergeant J & Swanson J (2007). A twin study of Halperin JM, Trampush JW, Miller CJ, Marks DJ & Newcorn JH (2008). Neuropsychological in children with autism spectrum disorder. Journal of the American attention-deficit/hyperactivity disorder dimensions rated by the strengths and weaknesses of ADHD-symptoms and normal-behavior (SWAN) scale. Biological Psychiatry, 61(5), 700-705. Hazell P (2007). Drug therapy for attention-deficit/hyperactivity disorder-like symptoms in autistic disorder. Journal of Paediatrics and Child Health, 43(1-2), 19-24. Hervey AS, Epstein JN, Curry JF, Tonev S, Arnold L, Conners CK, et al. (2006). Reaction time distribution analysis of neuropsychological performance in an ADHD sample. Child Neuropsychology, 12(2) 125-140. Hoekstra RA, Bartels M, Cath DC & Boomsma DI (2008). Factor structure, reliability and criterion validity of the Autism-Spectrum Quotient (AQ): a study in Dutch population and patient groups. Journal of Autism and Developmental Disorders, 38(8), 1555-1566. Holtmann M, Bolte S & Poustka F (2007). Attention deficit hyperactivity disorder symptoms in pervasive developmental disorders: association with autistic 240 241 Chapter 10 behavior domains and coexisting psychopathology. Psychopathology, Just MA, Keller TA, Malave VL, Kana RK & Varma S (2012). Autism as a neural 40, 172-177. systems disorder: a theory of frontal-posterior underconnectivity. Neuroscience & Biobehavioral Reviews, 36(4), 1292-1313. Hyman SE (2010). The diagnosis of mental disorders: the problem of reification. Annual Review of Clinical Psychology, 6, 155-179. Hyman SE (2007). Can neuroscience be integrated into the DSM-V? Nature Reviews Neuroscience, 8(9), 725-732. I Kasper LJ, Alderson RM & Hudec KL (2012). Moderators of working memory deficits in children with attention-deficit/hyperactivity disorder (ADHD): a meta-analytic review. Clinical Psychology Review, 32(7), 605-617. Kats-Gold I, Besser A & Priel B (2007). The role of simple emotion recognition Ivry RB (2003). Cerebellar involvement in clumsiness and other developmental disorders. Neural Plasticity, 10(1-2), 141-153. J skills among school aged boys at risk of ADHD. Journal of Abnormal Child Psychology, 35(3), 363-378. Kaufman AS, Balgopal R, Kaufman JC & McLean JE (1994). WISC-III Short Forms: Jaffee SR & Price TS (2007). Gene-environment correlations: a review of the evidence and implications for prevention of mental illness. Molecular Psychiatry, 12(5), 432-442. Johnson KA, Robertson IH, Kelly SP, Silk TJ, Barry E, Daibhis A, et al. (2007) Dissociation in performance of children with ADHD and high-functioning autism on a task of sustained attention. Neuropsychologia, 45, 22342245. de Jong CG, van de Voorde S, Roeyers H, Raymaekers R, Allen AJ, Knijff S, et al. (2009). Differential effects of atomoxetine on executive functioning and lexical decision in attention-deficit/hyperactivity disorder and reading disorder. Journal of Child and Adolescent Psychopharmacology, 19(6), 699-707. Jou RJ, Handen BL & Hardan AY (2005). Retrospective assessment of atomoxetine in children and adolescents with pervasive developmental disorders. Journal of Child and Adolescent Psychopharmacology, 15(2), 325-330. 242 K Psychometric Properties vs. Clincal Relevance vs. Practical Utility. Paper presented at the Annual Meeting of the Mid-South Educational Research Association. Nashville, TN, November 9-11. Kendler & Neale (2010) Endophenotype: a conceptual analysis. Molecular Psychiatry, 15(8), 789-797 Kim YS, Leventhal BL, Koh YJ, Fombonne E, Laska E, Lim EC, et al. (2011). Prevalence of autism spectrum disorders in a total population sample. American Journal of Psychiatry, 168, 904-912. Klein C, Wendling K, Huettner P, Ruder H& Peper M (2006). Intra-subject variability in attention-deficit hyperactivity disorder. Biological Psychiatry, 60, 10881097. Knouse LE, Barkley RA & Murphy KR (2013). Does executive functioning (EF) predict depression in clinic-referred adults? EF tests vs. rating scales. Journal of Affective Disorders, 145(2), 270-275. 243 Chapter 10 Kofler MJ, Rapport MD, Sarver DE, Raiker JS, Orban SA, Friedman LM & Kolomeyer EG (2013). Reaction time variability in ADHD: A meta-analytic Le Couteur A, Lord C & Rutter M (2003). Autism Diagnostic Interview-Revised (ADI-R). Los Angeles, CA: Western Psychological Services. review of 319 studies. Clinical Psychology Review, 33(6), 795-811. Lee YS, Han DH, Lee JH & Choi TY (2010). The Effects of Methylphenidate on Kotte A, Joshi G, Fried R, Uchida M, Spencer A, Woodworth KY, et al.(2013). Neural Substrates Associated with Interference Suppression in Children Autistic Traits in Children With and Without ADHD. Pediatrics, 132(3), with ADHD: A Preliminary Study Using Event Related fMRI. Psychiatry 612-622. Investigation, 7(1), 49-54. Kramer MD, Krueger RF & Hicks BM (2008). The role of internalizing and externalizing liability factors in accounting for gender differences in the Lehman EL (1975). Nonparametrics: Statistical methods based on ranks. San Francisco, CA: Holden-Day. prevalence of common psychopathological syndromes. Psychological Medicine, 38(1), 51-61. Leonard MA, Milich R & Lorch EP (2011). The role of pragmatic language use in mediating the relation between hyperactivity and inattention and social Kuntsi J, Wood AC, Rijsdijk F, Johnson KA, Andreou P, Albrecht B, et al.(2010). Separation of cognitive impairments in attention-deficit/hyperactivity skills problems. Journal of Speech Language and Hearing Research, 54(2), 567-579. disorder into 2 familial factors. Archives of General Psychiatry, 67(11), 1159-1167. Leth-Steensen C, Elbaz ZK & Douglas VI (2000). Mean response times, variability, and skew in the responding of ADHD children: a response time Kuntsi J, Oosterlaan J & Stevenson J (2001). Psychological mechanisms in distributional approach. Acta Psychologica, 104(2), 167-190. hyperactivity: I. Response inhibition deficit, working memory impairment, delay aversion, or something else? Journal of Child Psychology and Psychiatry, 42(2), 199-210. Levy F, Hay DA, McStephen M, Wood C & Waldman I (1997) Attention deficit hyperactivity disorder: a category or a continuum? Genetic analysis of a large-scale twin study. Journal of the American Acadamy of Child & L Adolescent Psychiatry, 36(6), 737-744. Lakes KD, Swanson JM & Riggs M (2012). The reliability and validity of the English and Spanish Strengths and Weaknesses of ADHD and Normal behavior rating scales in a preschool sample: continuum measures of hyperactivity and inattention. Journal of Attenention Disorders, 16(6), 510-516. Larsson H, Anckarsater H, Rastam M, Chang Z & Lichtenstein P (2012). Childhood attention-deficit hyperactivity disorder as an extreme of a continuous Leyfer OT, Folstein SE, Bacalman S, Davis NO, Dinh E, Morgan J, et al. (2006). Comorbid psychiatric disorders in children with autism: interview development and rates of disorders. Journal of Autism and Developmental Disorders, 36(7), 849-861. Los SA (1996). On the origin of mixing costs: exploring information processing in pure and mixed blocks of trials. Acta Psychologica, 94, 145-188. trait: a quantitative genetic study of 8,500 twin pairs. Journal of Child Psychology and Psychiatry, 53(1), 73-80. 244 245 Chapter 10 Li JJ & Lee SS (2010). Latent class analysis of antisocial behavior: interaction of Lundstrom S, Chang Z, Rastam M, Gillberg C, Larsson H, Anckarsater H & serotonin transporter genotype and maltreatment. Journal of Abnormal Lichtenstein P (2012). Autism spectrum disorders and autistic like traits: Child Psychology, 38(6), 789-801. similar etiology in the extreme end and the normal variation. Archives of General Psychiatry, 69(1), 46-52. Li JJ & Lee SS (2012). Interaction of dopamine transporter (DAT1) genotype and maltreatment for ADHD: a latent class analysis. Journal of Child Psychology and Psychiatry, 53(9), 997-1005. Lichtenstein P, Carlstrom E, Rastam M,Gillberg C & Anckarsater H (2010). The genetics of autism spectrum disorders and related neuropsychiatric disorders in childhood. American Journal of Psychiatry, 167(11), 13571363. van Lieshout M, Luman M, Buitelaar J, Rommelse NNJ & Oosterlaan J (2013). Does neurocognitive functioning predict future or persistence of ADHD? A systematic review. Clinical Psychology Review, 33(4) 539–560. Lijffijt M, Kenemans JL, Verbaten MN & van Engeland H (2005). A meta-analytic review of stopping performance in attention-deficit/hyperactivity disorder: deficient inhibitory motor control? Journal of Abnormal Psychology, 114(2), 216-222. Liu Y, Cherkassky VL, Minshew NJ & Just MA (2011). Autonomy of lower-level perception from global processing in autism: evidence from brain activation and functional connectivity. Neuropsychologia, 49(7), 21052111. Lord C & Jones RM (2012). Annual research review: re-thinking the classification of autism spectrum disorders. Journal of Child Psychology and Psychiatry, 53(5), 490-509. Loveland KA, Tunali-Kotoski B, Chen YR, Ortegon J, Pearson DA, Brelsford KA & Gibbs MC (1997). Emotion recognition in autism: verbal and nonverbal information. Developmental Psychopathology, 9, 579-593. 246 M Maher B (2008). Personal genomes: The case of the missing heritability. Nature, 456 (7218), 18-21. Maister L & Plaisted-Grant KC (2011). Time perception and its relationship to memory in Autism Spectrum Conditions. Developmental Science, 14(6), 1311-1322. Marcus DK & Barry TD (2011). Does attention-deficit/hyperactivity disorder have a dimensional latent structure? A taxometric analysis. Journal of Abnormal Psychology, 120(2), 427-442. Martin J, Hamshere ML, O’Donovan MC, Rutter M & Thapar A (2014). Factor structure of autistic traits in children with ADHD. Journal of Autism and Developmental Disorders, 44(1), 204-215. Martin JS, Poirier M & Bowler DM (2010). Brief report: Impaired temporal reproduction performance in adults with autism spectrum disorder. Journal of Autism and Developmental Disorders, 40(5), 640-646. Marx I, Hubner T, Herpertz SC, Berger C, Reuter E, Kircher T, et al. (2010). Crosssectional evaluation of cognitive functioning in children, adolescents and young adults with ADHD. Journal of Neural Transmission, 117(3), 403419. McAuley T, Crosbie J, Charach A & Schachar R (2013). The persistence of cognitive deficits in remitted and unremitted ADHD: a case for the stateindependence of response inhibition. J Child Psychology and Psychiatry, 55(3), 292-300 247 Chapter 10 McCarthy MI, Abecasis GR, Cardon LR, Goldstein DB, Little J, Ioannidis JP & Miller G (2010). Psychiatry. Beyond DSM: seeking a brain-based classification of Hirschhorn JN (2008). Genome-wide association studies for complex mental illness. Science, 327(5972), 1437. traits: consensus, uncertainty and challenges. Nature Reviews Genetics, 9(5), 356-369. Ming X & Walters AS (2009). Autism spectrum disorders, attention deficit/ hyperactivity disorder, and sleep disorders. Current Opinion in Pulmonary McCutcheon AL (1987) Latent Class Analysis. Beverly Hills, CA: Sage. Medicine, 15(6), 578-584. McGrath J, Johnson K, Ecker C, O’Hanlon E, Gill M, Gallagher L & Garavan van Mourik R, Oosterlaan J & Sergeant JA (2005). The Stroop revisited: a meta- H (2012). Atypical visuospatial processing in autism: insights from analysis of interference control in AD/HD. Journal of Child Psychology functional connectivity analysis. Autism Research, 5(5), 314-330. and Psychiatry, 46(2), 150-165. van Meel CS, Oosterlaan J, Heslenfeld DJ & Sergeant JA (2005). Motivational Mulligan A, Anney RJ, O’Regan M, Chen W, Butler L, Fitzgerald M, et al. (2009). effects on motor timing in attention-deficit/hyperactivity disorder. Journal Autism symptoms in attention-deficit/hyperactivity disorder: a familial of the American Academy of Child and Adolescent Psychiatry, 44(5), 451- trait which correlates with conduct, oppositional defiant, language and 460. motor disorders. Journal of Autism and Developmental Disorders, 39(2), 197-209. van der Meer JMJ, Lappenschaar MGA, Hartman CA, Greven CU, Buitelaar JK & Rommelse NNJ (in press) Homogeneous combinations of ASD-ADHD Munson J, Dawson G, Sterling L, Beauchaine T, Zhou A, Elizabeth K, et al. (2008). traits and their cognitive and behavioral correlates in a population-based Evidence for latent classes of IQ in young children with autism spectrum sample. Journal of Attention Disorders. disorder. American Journal of Mental Retardation, 113(6), 439-452. van der Meer JMJ, Oerlemans AM, van Steijn DJ, Lappenschaar MG, de Sonneville Murray MJ (2010). Attention-deficit/Hyperactivity Disorder in the context of Autism LM, Buitelaar JK & Rommelse NNJ (2012). Are autism spectrum disorder spectrum disorders. Current Psychiatry Reports, 12(5), 382-388. and attention-deficit/hyperactivity disorder different manifestations of one overarching disorder? Cognitive and symptom evidence from a Muthén LK & Muthén BO (2006). Mplus User’s Guide. Statistical Analysiswith Latent Variables, 6thEd., version 6.1: Los Angeles. clinical and population-based sample. Journal of the American Academy of Child and Adolescent Psychiatry, 51(11), 1160-1172. N Millan MJ, Agid Y, Brune M, Bullmore ET, Carter CS, Clayton NS, et al.(2012). Cognitive dysfunction in psychiatric disorders: characteristics, causes and the quest for improved therapy. Nature Reviews Drug Discovery, 11(2), 141-168. 248 Nandam LS, Hester R, Wagner J, Cummins TD, Garner K, Dean AJ, et al.(2011). Methylphenidate but not atomoxetine or citalopram modulates inhibitory control and response time variability. Biological Psychiatry, 69(9), 902904. 249 Chapter 10 Neale MC & Kendler KS (1995). Models of comorbidity for multifactorial disorders. Nijmeijer JS, Arias-Vasquez A, Rommelse NN, Altink ME, Anney RJ, Asherson P, American Journal of Human Genetics, 57(4), 935-953. et al. (2010). Identifying loci for the overlap between attention-deficit/ hyperactivity disorder and autism spectrum disorder using a genome- Nederlands Vereniging voor Psychiatrie (NVvP) (2009). Richtlijn diagnostiek en wide QTL linkage approach. Journal of the American Academy of Child behandeling autismespectrumstoornissen bij kinderen en jeugdigen Nettle D (2006). The evolution of personality variation in humans and other and Adolescent Psychiatry, 49(7), 675-685. Nijmeijer JS, Minderaa RB, Buitelaar JK, Mulligan A, Hartman CA & Hoekstra PJ animals. The American Psychologist 61(6), 622-631. (2008). Attention-deficit/hyperactivity disorder and social dysfunctioning. Clinical Psychology Review, 28, 692-708. Neuman RJ, Sitdhiraksa N, Reich W, Ji TH, Joyner CA, Sun LW & Todd RD (2005). Estimation of prevalence of DSM-IV and latent class-defined ADHD Nikolas MA & Burt SA (2010). Genetic and environmental influences on ADHD subtypes in a population-based sample of child and adolescent twins. symptom dimensions of inattention and hyperactivity: a meta-analysis. Twin Research & Human Genetics 8(4), 392-401. Journal of Abnormal Psychology, 119(1), 1-17. Newman ME (2006). Modularity and community structure in networks. Noreika V, Falter CM & Rubia K (2013). Timing deficits in attention-deficit/ Proceedings of the National Academy of Sciences of the United States of hyperactivity disorder (ADHD): evidence from neurocognitive and America, 103(23), 8577-8582. neuroimaging studies. Neuropsychologia, 51(2), 235-266. Nigg JT, Willcutt EG, Doyle AE & Sonuga-Barke EJ (2005). Causal heterogeneity in Nydén A, Niklasson L, Stahlberg O, Anckarsater H, Wentz E, Rastam M & attention-deficit/hyperactivity disorder: do we need neuropsychologically Gillberg C (2010). Adults with autism spectrum disorders and ADHD impaired subtypes? Biological Psychiatry, 57(11), 1224-1230. neuropsychological aspects. Research In Developmental Disabilities, 31(6), 1659-1668. Nigg JT, Goldsmith HH & Sachek J (2004). Temperament and attention deficit hyperactivity disorder: the development of a multiple pathway model. Nylund KL, Asparouhov T & Muthén B (2007). Deciding on the number of classes Journal of Clinical Child and Adolescent Psychology, 33(1), 42-53. in latent class analysis and growth mixture modeling: a Monte Carlo simulation study. Structural Equation Modeling, 14(4), 535-569. Nigg JT, Blaskey LG, Huang-Pollock CL & Rappley MD (2002). Neuropsychological executive functions and DSM-IV ADHD subtypes. Journal of the American Academy of Child and Adolescent Psychiatry, 41(1), 59-66. Nigg JT (1999). The ADHD response-inhibition deficit as measured by the stop task: replication with DSM-IV combined type, extension, and qualification. Journal of Abnormal Child Psychology, 27(5), 393-402. 250 O Oerlemans AM, Droste K, van Steijn DJ, de Sonneville LM, Buitelaar JK & Rommelse NN (2013) Co-segregation of social cognition, executive function and local processing style in children with ASD, their siblings and normal controls. Journal of Autism and Developmental Disorders, 43(12), 2764-2778. 251 Chapter 10 van der Oord S, Geurts HM, Prins PJ, Emmelkamp PM & Oosterlaan J (2012). Prepotent response inhibition predicts treatment outcome in attention Plomin R, Haworth CM & Davis OS (2009). Common disorders are quantitative traits. Nature Reviews Genetics, 10(12), 872-878. deficit/hyperactivity disorder. Child Neuropsychology, 18(1), 50-61. Plomin R (2013). Child development and molecular genetics: 14 years later. Child Oosterlaan J, Logan GD & Sergeant JA (1998). Response inhibition in AD/HD, CD, Development, 84, 104-120. comorbid AD/HD + CD, anxious, and control children: a meta-analysis of studies with the stop task. Journal of Child Psychology and Psychiatry, 39(3), 411-425. Polanczyk G, de Lima MS, Horta BL, Biederman J & Rohde LA (2007). The worldwide prevalence of ADHD: a systematic review and metaregression analysis. American Journal of Psychiatry, 164(6), 942-948. P Paterson NE, Ricciardi J, Wetzler C & Hanania T (2011). Sub-optimal performance in the 5-choice serial reaction time task in rats was sensitive to methylphenidate, atomoxetine and d-amphetamine, but unaffected by the COMT inhibitor tolcapone. Neuroscience Research, 69(1), 41-50. Pattij T, Schetters D, Schoffelmeer AN & van Gaalen MM (2012). On the improvement of inhibitory response control and visuospatial attention by indirect and direct adrenoceptor agonists. Psychopharmacology, 219(2), 327-340. Pauli-Pott U, Dalir S, Mingebach T, Roller A & Becker K (2013). Attention deficit/ hyperactivity and comorbid symptoms in preschoolers: Differences between subgroups in neuropsychological basic deficits. Child Neuropsychology, 20(2), 230-244. Pennington BF & Ozonoff S (1996). Executive functions and developmental psychopathology. Journal of Child Psychology and Psychiatry, 37(1), 5187. Polderman TJ, Derks EM, Hudziak JJ, Verhulst FC, Posthuma D & Boomsma DI (2007). Across the continuum of attention skills: a twin study of the SWAN ADHD rating scale. Journal of Child Psychology and Psychiatry, 48(11), 1080-1087. Posey DJ, Aman MG, McCracken JT, Scahill L, Tierney E, Arnold LE, et al. (2007). Positive effects of methylphenidate on inattention and hyperactivity in pervasive developmental disorders: an analysis of secondary measures. Biological Psychiatry, 61(4), 538-544. Posey DJ, Wiegand RE, Wilkerson J, Maynard M, Stigler KA & McDougle CJ (2006). Open-label atomoxetine for attention-deficit/ hyperactivity disorder symptoms associated with high-functioning pervasive developmental disorders. Journal of Child and Adolescent Psychopharmacology, 16(5), 599-610. Preacher KJ, Rucker DD, MacCallum RC & Nicewander WA (2005) Use of the extreme groups approach: A critical reexamination and new recommendations. Psychological Methods, 10(2), 178-192. Pellicano E, Maybery M, Durkin K & Maley A (2006). Multiple cognitive capabilities/ deficits in children with an autism spectrum disorder: “weak” central coherence and its relationship to theory of mind and executive control. Developmental Psychopathology, 18(1), 77-98. 252 253 Chapter 10 R Research Units On Pediatric Psychopharmacology (RUPP) Autism Network (2012). Rajendran K, Rindskopf D, O’Neill S, Marks DJ, Nomura Y &Halperin JM (2013). Neuropsychological functioning and severity of ADHD in early childhood: A four-year cross-lagged study. Journal of Abnormal Psychology, 122(4), 1179-1188. Rajendran K, Trampush JW, Rindskopf D, Marks DJ, O’Neill S & Halperin JM (2013). Association between variation in neuropsychological development and trajectory of ADHD severity in early childhood. American Journal of Psychiatry, 170(10), 1205-1211. Reiersen AM & Todorov AA (2013). Exploration of ADHD Subtype Definitions and Co-Occurring Psychopathology in a Missouri Population-Based Large Sibship Sample. Scandinavian Journal of Child and Adolescent Psychiatry and Psychology, 1, 3-13 Reiersen, AM, Constantino JN & Todd RD (2008). Co-occurrence of motor problems and autistic symptoms in attention-deficit/hyperactivity disorder. Journal of the American Academy of Child and Adolescent Psychiatry, 47(6), 662-672. Reiersen AM & Todd RD (2008). Co-occurrence of ADHD and autism spectrum disorders: phenomenology and treatment. Expert Review of Neurotherapeutics, 8(4), 657-669. Reiersen AM, Constantino JN, Grimmer M, Martin NG & Todd RD (2008). Evidence for shared genetic influences on self-reported ADHD and autistic symptoms in young adult Australian twins. Twin Research and Human Genetics, 11(6), 579-585. Reiersen AM, Constantino JN, Volk HE & Todd RD (2007). Autistic traits in a population-based ADHD twin sample. Journal of Child Psychology and Psychiatry, 48(5), 464-472. 254 Effects of risperidone and parent training on adaptive functioning in children with pervasive developmental disorders and serious behavioral problems. Journal of the American Academy of Child and Adolescent Psychiatry, 51(2), 136-146. Research Units On Pediatric Psychopharmacology (RUPP) Autism Network (2009). Medication and parent training in children with pervasive developmental disorders and serious behavior problems: results from a randomized clinical trial. Journal of the American Academy of Child and Adolescent Psychiatry, 48(12), 1143-1154 Research Units on Pediatric Psychopharmacology (RUPP) Autism Network (2005). Randomized, controlled, crossover trial of methylphenidate in pervasive developmental disorders with hyperactivity. Archives of General Psychiatry, 62, 1266-1274. Richmond LL, Thorpe M, Berryhill ME, Klugman J & Olson IR (2013). Individual differences in autistic trait load in the general population predict visual working memory performance. Quarterly Journal of Experimental Psychology, 66(6), 1182-1195. Robinson EB, Koenen KC, McCormick MC, Munir K, Hallett V, Happé F, et al. (2011). Evidence that autistic traits show the same etiology in the general population and at the quantitative extremes (5%, 2.5%, and 1%). Archives of General Psychiatry, 68(11), 1113-1121. Robinson EB, Munir K, Munafo MR, Hughes M, McCormick MC & Koenen KC (2011). Stability of autistic traits in the general population: further evidence for a continuum of impairment. Journal of the American Academy of Child and Adolescent Psychiatry, 50(4), 376-384. Robinson ES, Eagle DM, Mar AC, Bari A, Banerjee G, Jiang X, et al.(2008). Similar effects of the selective noradrenaline reuptake inhibitor atomoxetine on 255 Chapter 10 three distinct forms of impulsivity in the rat. Neuropsychopharmacology, 33(5), 1028-1037. Rubia K, Halari R, Cubillo A, Smith AB, Mohammad AM, Brammer M & Taylor E (2011). Methylphenidate normalizes fronto-striatal underactivation during interference inhibition in medication-naive boys with attention- Rommelse NN, Geurts HM, Franke B, Buitelaar JK & Hartman CA (2011) A review on cognitive and brain endophenotypes that may be common in autism deficit hyperactivity disorder. Neuropsychopharmacology, 36(8), 15751586. spectrum disorder and attention-deficit/hyperactivity disorder and facilitate the search for pleiotropic genes. Neuroscience & Biobehavioral Reviews, 35(6), 1363-1396. Rubia K, Halari R, Christakou A & Taylor E (2009). Impulsiveness as a timing disturbance: neurocognitive abnormalities in attention-deficit hyperactivity disorder during temporal processes and normalization Rommelse NN, Franke B, Geurts HM, Hartman CA & Buitelaar JK (2010) Shared heritability of attention-deficit/hyperactivity disorder and autism spectrum with methylphenidate. Philosophical Transactions of the RoyalSociety B: Biological Sciences, 364(1525), 1919-1931. disorder. European Child and Adolescent Psychiatry, 19(3), 281-295. Rubia K, Smith AB, Brammer MJ, Toone B & Taylor E (2005). Abnormal brain Rommelse NN, Altink ME, Oosterlaan J, Beem L, Buschgens CJ, Buitelaar J activation during inhibition and error detection in medication-naive & Sergeant JA (2008). Speed, variability, and timing of motor output adolescents with ADHD. American Journal of Psychiatry, 162(6), 1067- in ADHD: which measures are useful for endophenotypic research? 1075. Behavior Genetics, 38(2), 121-132. Rubia K, Overmeyer S, Taylor E, Brammer M, Williams SC, Simmons A & Bullmore Rommelse NN, Altink ME, de Sonneville LM, Buschgens CJ, Buitelaar J, ET (1999). Hypofrontality in attention deficit hyperactivity disorder during Oosterlaan J & Sergeant JA (2007). Are motor inhibition and cognitive higher-order motor control: a study with functional MRI. American Journal flexibility dead ends in ADHD? Journal of Abnormal Child Psychology, of Psychiatry, 156(6), 891-896. 35, 957-967. Rucklidge JJ & Tannock R (2002). Neuropsychological profiles of adolescents Ronald A, Edelson LR, Asherson P & Saudino KJ (2010). Exploring the relationship between autistic-like traits and ADHD behaviors in early childhood: with ADHD: effects of reading difficulties and gender. Journal of Child Psychology and Psychiatry, 43(8), 988-1003. findings from a community twin study of 2-year-olds. Journal of Abnormal Child Psychology, 38(2), 185-196. Rudie JD, Brown JA, Beck-Pancer D, Hernandez LM, Dennis EL, Thompson PM, et al. (2013). Altered functional and structural brain network organization Ronald A, Simonoff E, Kuntsi J, Asherson P & Plomin R (2008). Evidence for in autism. Neuroimage Clinical, 2, 79-94. overlapping genetic influences on autistic and ADHD behaviours in a community twin sample. Journal of Child Psychology and Psychiatry and Allied Disciplines, 49(5), 535-542. Rutter M, Bailey A, Berument S, Lecouteur A, Lord C & Pickles A (2003). Social Communication Questionnaire (SCQ) Los Angeles, CA: Western Psychological Services. 256 257 Chapter 10 S disorder, combined type: functional magnetic resonance imaging study. Santosh PJ, Baird G, Pityaratstian N, Tavare E & Gringras P (2006). Impact of comorbid autism spectrum disorders on stimulant response in children with attention deficit hyperactivity disorder: a retrospective and prospective effectiveness study. Child: Care, Health and Development, 32(5), 575-583. Schachar R, Mota VL, Logan GD, Tannock R & Klim P (2000). Confirmation of an inhibitory control deficit in attention-deficit/hyperactivity disorder. Journal of Abnormal Child Psychology, 28(3), 227-235. Scheres A, OosterlaanJ, Geurts H, Morein-Zamir S, Meiran N, Schut H, et al. (2004). Executive functioning in boys with ADHD: primarily an inhibition deficit? Archives of Clinical Neuropsychology, 19(4), 569-594. Scheres A, Oosterlaan J, Swanson J, Morein-Zamir S, Meiran N, Schut H, et al. (2003) The effect of methylphenidate on three forms of response inhibition in boys with AD/HD. Journal of Abnormal Child Psychology, 31(1), 105-120. Schwartz K & Verhaeghen P (2008). ADHD and Stroop interference from age 9 to age 41 years: a meta-analysis of developmental effects. Psychological Medicine, 38(11), 1607-1616. Shah A & Frith U (1983). An islet of ability in autistic children: a research note. Journal of Child Psychology and Psychiatry, 24(4), 613-620. Sikora DM, Vora P, Coury DL & Rosenberg D (2012). Attention-deficit/hyperactivity disorder symptoms, adaptive functioning, and quality of life in children with autism spectrum disorder. Pediatrics, 130(2), S91-97. Silk T, Vance A, Rinehart N, Egan G, O’Boyle M, Bradshaw JL & Cunnington British Journal of Psychiatry, 187, 282-283. Simonoff E, Pickles A, Charman T, Chandler S, Loucas T & Baird G (2008). Psychiatric disorders in children with autism spectrum disorders: prevalence, comorbidity, and associated factors in a population-derived sample. Journal of the American Academy of Child and Adolescent Psychiatry, 47(8), 921-929. Sinzig J, Morsch D, Bruning N, Schmidt MH & Lehmkuhl G (2008). Inhibition, flexibility, working memory and planning in autism spectrum disorders with and without comorbid ADHD-symptoms. Child and Adolescent Psychiatry and Mental Health, 2, 4. de Sonneville LMJ (2005). Amsterdam Neuropsychological Tasks: scientific and clinical applications. Tijdschrift voor Neuropsychologie, 1, 27-41. de Sonneville LMJ (1999). Amsterdam Neuropsychological Tasks; a computeraided assessment program. Cognitive ergonomics, clinical assessment, and computer-assisted learning: Computers in Psychology, 6, 187-203. Sonuga-Barke EJ (2005). Causal models of attention-deficit/hyperactivity disorder: from common simple deficits to multiple developmental pathways. Biological Psychiatry, 57(11), 1231-1238. Sonuga-Barke EJ (2002). Psychological heterogeneity in AD/HD- a dual pathway model of behaviour and cognition. Behavioral Brain Research, 130(1-2), 29-36. Spencer T, Biederman J, Wilens T, Prince J, Hatch M, Jones J, et al. (1998). Effectiveness and tolerability of tomoxetine in adults with attention deficit hyperactivity disorder. American Journal of Psychiatry, 155(5), 693-695. R (2005). Fronto-parietal activation in attention-deficit hyperactivity 258 259 Chapter 10 Spencer TJ, Faraone SV, Michelson D, Adler LA, Reimherr FW, Glatt SJ & Biederman Szelag E, Kowalska J, Galkowski T & Poppel E(2004). Temporal processing J (2006). Atomoxetine and adult attention-deficit/hyperactivity disorder: deficits in high-functioning children with autism. British Journal of the effects of comorbidity. Journal of Clinical Psychiatry, 67(3), 415-420. Psychology, 95(3) 269-282. Spiker D, Lotspeich LJ, Dimiceli S, Myers RM & Risch N. (2002). Behavioral phenotypic variation in autism multiplex families: evidence for a continuous severity gradient. American Journal of Medical Genetics, 114(2), 129-136. of attention deficit hyperactivity disorder and autistic spectrum disorder. Applied Clinical Genetics, 2, 7-13. (2012). ADHD and autism: differential diagnosis or overlapping traits? A B, et al. (2012). The co-occurrence of Attention-Deficit/Hyperactivity Disorder and Autism Spectrum Disorder symptoms in parents of children with ASD or ASD with ADHD. Journal of Child Psychology and Psychiatry, 53(9), 954–963. Stigler KA, Desmond LA, Posey DJ, Wiegand RE & McDougle CJ (2004). A naturalistic retrospective analysis of psychostimulants in pervasive disorders. Journal Taylor MJ, Charman T, Robinson EB, Plomin R, Happé F, Asherson P & Ronald A (2013). Developmental associations between traits of autism spectrum disorder and attention deficit hyperactivity disorder: a genetically van Steijn DJ, Richards JS, Oerlemans AM, De Ruiter SW, Van Aken MAG, Franke of Child and Adolescent Psychopharmacology, 14(1), 49-56. St. Pourcain B, Mandy WP, Heron J, Golding J, Smith GD & Skuse DH (2011). Links between co-occurring social-communication and hyperactiveinattentive trait trajectories. Journal of the American Academy of Child & Adolescent Psychiatry, 50(9), 892-902. St. Pourcain B, Wang K, Glessner JT, Golding J, Steer C, Ring SM, et al. (2010). Association between a high-risk autism locus on 5p14 and social communication spectrum phenotypes in the general population. American Journal of Psychiatry, 167(11), 1364-1372. 260 Taurines R, Schwenck C, Westerwald E, Sachse M, Siniatchkin M & Freitag C selective review. Attention Deficit Hyperactivity Disorder, 4(3), 115-139. Stam AJ, Schothorst PF, Vorstman JA & Staal WG (2009). The genetic overlap developmental T informative, longitudinal twin study. Psychological Medicine, 43(8), 1735-1746. Taylor E, Sandberg S, Thorley G & Giles S (1991). The epidemiology of childhood hyperactivity. Oxford University Press, New York. Thissen A, Luman M, Hartman C, Hoekstra P, van Lieshout M, Franke B, et al. (under revision). ADHD and motor timing in adolescents and their parents: endophenotype characteristics of reaction time variability vary with age. Journal of the American Academy of Child and Adolescent Psychiatry. Thompson L, Thompson M & Reid A (2010). Neurofeedback outcomes in clients with Asperger’s syndrome. Applied Psychophysiology and Biofeedback, 35(1), 63-81. Todd RD, Sitdhiraksa N, Reich W, Ji TH, Joyner CA, Heath AC & Neuman RJ (2002). Discrimination of DSM-IV and latent class attention-deficit/hyperactivity disorder subtypes by educational and cognitive performance in a population-based sample of child and adolescent twins. Journal of the American Academy of Child and Adolescent Psychiatry 41(7), 820-828. 261 Chapter 10 Toplak ME, Dockstader C & Tannock R (2006). Temporal information processing hyperactivity disorder, combined type: a functional MRI study. Molecular in ADHD: findings to date and new methods. Journal of Neuroscience Psychiatry, 12(9), 826-832. Methods, 151(1), 15-29. Vaughn AJ, Epstein JN, Rausch J, Altaye M, Langberg J, Newcorn JH, et Tourangeau R, Rips L & Rasinki K (2000). The Psychology of Survey Response. al.(2011). Relation between outcomes on a continuous performance test New York: Cambridge University Press. and ADHD symptoms over time. Journal of Abnormal Child Psychology, 39(6), 853-864. Trimbosinstituut (2005) Multidisciplinaire richtlijn ADHD, Richtlijn voor de diagnostiek en behandeling van ADHD bij kinderen en jeugdigen. Veatch OJ, Veenstra-Vanderweele J, Potter M, Pericak-Vance MA & Haines JL (2014). Genetically Meaningful Phenotypic Subgroups in Autism Troost PW, Steenhuis MP, Tuynman-Qua HG, Kalverdijk LJ, Buitelaar JK, Minderaa Spectrum Disorders. Genes, Brain and Behavior, 13(3), 276-285. RB & Hoekstra PJ (2006). Atomoxetine for attention-deficit/hyperactivity disorder symptoms in children with pervasive developmental disorders: Verté S, Geurts HM, Roeyers H, Oosterlaan J & Sergeant JA (2006). The a pilot study. Journal of Child and Adolescent Psychopharmacology, relationship of working memory, inhibition, and response variability in 16(5), 611-619. child psychopathology. Journal of Neuroscience Methods, 151, 5-14. Turner DC, Clark L, Dowson J, Robbins TW & Sahakian BJ (2004). Modafinil Vissers ME, Cohen MX & Geurts HM (2012). Brain connectivity and high improves cognition and response inhibition in adult attention-deficit/ functioning autism: a promising path of research that needs refined hyperactivity disorder. Biological Psychiatry, 55(10), 1031-1040. models, methodological convergence, and stronger behavioral links. Neuroscience and Biobehavioral Reviews, 36(1), 604-625. U Uher R & Rutter M (2012). Basing psychiatric classification on scientific foundation: Volk HE, Todorov AA, Hay DA & Todd RD (2009). Simple identification of complex ADHD subtypes using current symptom counts. Journal of the American problems and prospects. International Review of Psychiatry, 24(6), 591- Academy of Child & Adolescent Psychiatry, 48(4), 441-450. 605. W V Vaidya CJ, Bunge SA, Dudukovic NM, Zalecki CA, Elliott GR & Gabrieli JD (2005). Altered neural substrates of cognitive control in childhood ADHD: evidence from functional magnetic resonance imaging. American Journal of Psychiatry, 162(9), 1605-1613. Vance A, Silk TJ, Casey M, Rinehart NJ, Bradshaw JL, Bellgrove MA & Cunnington Walters JT & Owen MJ (2007). Endophenotypes in psychiatric genetics. Molecular Psychiatry, 12(10), 886-890. Wang L, Zhu C, He Y, Zang Y, Cao Q, Zhang H, et al. (2009). Altered small-world brain functional networks in children with attention-deficit/hyperactivity disorder. Human Brain Mapping, 30(2), 638-649. R (2007). Right parietal dysfunction in children with attention deficit 262 263 Chapter 10 Wass S (2011). Distortions and disconnections: disrupted brain connectivity in autism. Brain and Cognition, 75(1), 18-28. Zeiner P, Gjevik E & Weidle B (2011). Response to atomoxetine in boys with highfunctioning autism spectrum disorders and attention deficit/hyperactivity disorder. Acta Paediatrica, 100(9), 1258-1261. Wechsler D (2002). WISC-III Handleiding. London: Psychological Corporation. Wechsler D (2000). WAIS-III Nederlandstalige Bewerking. Technische handleiding. London: Psychological Corporation. Wechsler D (1989). Wechsler Preschool and Primary Scale of IntelligenceRevised. San Antonio, TX: Psychological Corporation. Widiger TA & Samuel DB (2005). Diagnostic categories or dimensions? A question for the Diagnostic and Statistical Manual of Mental Disorders-fifth edition. Journal of Abnormal Child Psychology, 114, 494-504 Willcutt EG, Nigg JT, Pennington BF, Solanto MV, Rohde LA, Tannock R, et al.(2012). Validity of DSM-IV Attention Deficit/Hyperactivity Disorder Symptom Dimensions and Subtypes. Journal of Abnormal Psychology, 121(4), 991-1010. Willcutt EG (2005). The etiology of ADHD: Behavioural and molecular genetic approaches. Cognitive and affective neuroscience of psychopathology. Oxford University Press, Oxford Y Yang L, Cao Q, Shuai L, Li H, Chan RC & Wang Y (2011). Comparative study of OROS-MPH and atomoxetine on executive function improvement in ADHD: a randomized controlled trial. International Journal of Neuropsychopharmacology, 15, 1-12. Z de Zeeuw P, Weusten J, van Dijk S, van Belle J & Durston S (2012). Deficits in cognitive control, timing and reward sensitivity appear to be dissociable in ADHD. PLoS One, 7(12). 264 265 Acknowledgements in Dutch 266 Het zit er op. Wauw. Chapter 11 Het promotietraject heb ik als een avontuur ervaren. Ik werd omringd door ijzersterke experts die me stimuleerden om veel te leren & fijne collega’s die ervoor zorgden dat ik ondertussen niet sociaal achterop raakte. Perfect. Ik houd er van om hoge Acknowledgements in Dutch doelen na te streven en hard te werken. Altijd al gedaan. Een strebertje to the limit? Misschien wel. Ik had mijzelf ten doel gesteld om met mijn promotieonderzoek toch minimaal tot zorgverbetering voor de kinder-en jeugdpsychiatrie te komen. Daarvoor moest gedegen wetenschappelijk onderzoek bedreven worden. Met die missie in gedachten ben ik het promotietraject, al was het bij vlagen ietwat neurotisch, goed doorgekomen. Er zijn veel mensen die een cruciale bijdrage hebben geleverd aan de totstand koming van dit proefschrift, enkelen van hen wil ik hieronder expliciet noemen. Mijn promotor prof. dr. Buitelaar. Jan, uiteraard heb ik grote bewondering voor je expertise en uiteraard heeft jouw feedback mijn papers verbeterd. Dat zegt alleen nog weinig over de verstandhouding tussen Jolanda-de-junior-onderzoeker en de Jan-de-promotor. Ik denk met veel plezier terug aan de overleggen met onze Israëlische vrienden, het stijldansen op een personeelsfeest (dat moet er toch wonderlijk uit hebben gezien :)), en je lachsalvo’s om toch wel voornamelijk je eigen grappen. Dat jij op afstand meestuurde en meedacht in dit promotietraject heeft het tot een succes gemaakt; bedankt voor het vertrouwen dat je in me gesteld hebt. Mijn eerste co-promotor dr. Lambregts-Rommelse. Nanda, ik heb het getroffen met je. Van meet af aan zaten we op één lijn. Qua ambities, qua werkstijl, qua persoonlijkheid. Onze overleggen gingen ook altijd even over de belangrijke zaken die zich buiten het werk om afspeelden. Ik vind het tof dat we zo’n vriendschappelijk contact hebben ontwikkeld. Werkinhoudelijk ben ik je dankbaar 269 Chapter 11 voor al je zinnige, vlotte en opbouwende kritieken. Meer dan eens mailde je ‘Dit van deze ouders en kinderen zijn we meer te weten gekomen over de relatie kunnen we hoog wegzetten, moeten we even 100% perfect opschrijven’. Zonder tussen gedrag en cognitieve prestaties. Met deze kennis begrijpen we weer iets jou was dit proefschrift er niet geweest. Period. Ik bewonder niet alleen hoe jij beter waardoor het ene kind ASS of ADHD ontwikkelt en het andere kind niet. tien promovendi op de rit weet te houden, hoe jij drie kleintjes thuis weet groot Allen hartelijk dank voor jullie inzet! te brengen, of hoe jij dat weet te combineren met de GZ-opleiding. Ik bewonder vooral hoe je dat allemaal met oprechte aandacht voor iedereen en een niet uitvoeren, hiervoor zijn we veel dank verschuldigd aan enkele leerkrachten en aflatend enthousiasme kunt. Dankjewel! directies. Onmisbaar was de betrokkenheid van De Hazesprong, De Bloemberg Veel kinderen mochten de cognitieve taken onder schooltijd en op school en Klein Heyendaal in Nijmegen, de Alfonsusschool te Enschede, Sinte Maerte in Mijn tweede co-promotor, dr. Hartman. Catharina, jouw inhoudelijke adviezen Breda, Gerardus Majella in Groesbeek, De Regenboog in Malden en Floriande en hebben de laatste papers naar een hoger niveau getild, terwijl je persoonlijke de 1e Montessorischool te Hoofddorp. begeleiding het laatste jaar van mijn promotietraject de nodige luchtigheid heeft gegeven. Wanneer ik naar jouw smaak te ongeduldig was, wist je me daar hulp gehad van vele studenten. Anne, Annemie, Dominique, Gertie, Gosia, met een ‘Je bent al net zo verschrikkelijk als Nanda!’ tactvol op te attenderen. Imke, Jennie, Jill, Kim, Lenthe, Maartje, Marike, Marjolein, Marthe, Myrthe, Geregeld heb ik hartelijk gelachen om je gebrekkige inhibitievermogen en Nienke, Nina, Noortje, Renate en Véronique, ik heb er veel plezier aan beleefd bondig geformuleerde antwoorden. Je scherpe inzicht enerzijds en je spontaniteit om jullie te coachen en enkelen van jullie te begeleiden bij het schrijven van de anderzijds is een unieke combinatie in de academische wereld. Ik heb van beide mastherthesis. Het was iedere week weer een uitdaging om de planning rond geprofiteerd, veel dank daarvoor! te krijgen. Wie gaat er op welke dag, naar welke stad, welk kind testen? Hebben Gelukkig hoefde ik alle kinderen niet in mijn eentje te zien en heb ik ze dan geen pauze, gymles, open dag of jarige juf? De dataverzameling kon vlot Leden van de manuscriptcommissie prof. dr. Willemsen, prof. dr. Bekkering en verlopen doordat jullie zo gemotiveerd en flexibel waren. Dank jullie wel. Ik hoop prof. dr. Roeyers en leden van de promotiecommissie mw. prof. dr. Geurts, mw. dr. dat de opgedane ervaringen jullie helpen in jullie verdere carrière en hoor graag Polderman en dr. Staal, hartelijk dank voor het bestuderen van mijn proefschrift hoe het jullie vergaat! en voor het opponeren bij de verdediging op woensdag 3 september. Mrs. prof. dr. Simonoff, thank you for the time and effort spent on reading my dissertation Er zijn veel collega’s die op directe of indirecte wijze bijgedragen aan mijn and for your willingness to act as opponent of my dissertation on Wednesday the werkvreugde. Patricia, bedankt voor het aanwakkeren van mijn liefde voor 3rd of September. academisch onderzoek. Rutger-Jan, Dorith, Sascha, Dorine en Kina bedankt voor de geboden mogelijkheden om de feeling voor het klinisch werk niet te Dit proefschrift is gebaseerd op data uit neuropsychologisch onderzoek verliezen. Shireen, Brechje en alle andere KCK-leden, bedankt voor de hulp bij bij honderden kinderen en jongeren uit heel Nederland, een flinke stapel het organiseren van het Karakter symposium ‘De brug tussen wetenschappelijk gedragsvragenlijsten die ouders over hun kinderen hebben ingevuld en onderzoek en klinische praktijk’. Het is zo’n interessant en relevant onderwerp; diagnostische interviews die bij ouders zijn afgenomen. Dankzij de medewerking hopelijk zullen er nog vele symposia volgen! 270 271 Chapter 11 Barbara Franke en Kimm van Hulzen, het paper dat we samen beoogden en ontwikkelgesprekken, op andere dagen waren we beiden zo ijverig dat te schrijven over susceptibility genes voor ASS en ADHD is helaas nog niet af. Ik we nauwelijks een woord wisselden. Tegenpolen op menig maar blijkbaar hoop dat deze belangrijke studie met behulp van meer samples alsnog doorgang verwaarloosbaar vlak. Het maakt jou & mij een zowel singulier en buitenissig zal vinden. Martijn Lappenschaar, Marcel Zwiers en Rogier Donders bedankt voor als flagrant en apert duo. Het promotietraject heeft onze vriendschap werkelijk jullie onmisbare deskundigheid aangaande statistiek. Heel fijn dat mails met als verdiept en verrijkt, wat een briljante benefit. onderwerp ‘Een heel kort vraagje’ zo vlot werden beantwoord! Corina Greven, Gigi van de Loo, Leo de Sonneville, Marjolein Luman, Monika Althaus, Myriam Eindhovenaren, Hagenaren en Utrechters. Door me in de weekenden samen Harfterkamp en Pieter Hoekstra hartelijk dank voor de prettige samenwerking. met jullie met volstrekt andere zaken bezig te houden dan promoveren kon ik Enschedese crew, ten dele inmiddels rasechte Amsterdammers, weer opladen voor nieuwe werkweken. De ‘Hé Jo, hoe gaat het eigenlijk met je De onderzoeksafdeling bij Karakter wordt bestierd door een tiental ambitieuze verslagje / scriptie / rapportje / werkstukje?’ werkte bovendien zeer relativerend. vrouwen. Geen 9-tot-5 mentaliteit te bekennen. Heerlijk. Publicaties, presentaties Dankjulliewel! Marlijn en Ingeborg, ik koester onze vriendschap die al zoveel en gewonnen beurzen samen vieren, tegenvallende inclusie, eindeloze analyses doorstaan heeft. Moniek, wat had ik je graag nog hier bij ons gehad! en bergen feedback samen vervloeken; het is heel belangrijk voor me geweest. Anoek, Daphne, Jennifer, Karlijn, Kirsten, Leonie, Loes, Mireille, Mirjam, Saskia Marloes, Annemiek en Martijn; mijn dierbare zus, zusje en broertje. De vraag of ik en Yvette, heel veel dank. Ik hoop en verwacht dat de toekomst veel goeds voor ooit weer terugkeer naar Het Twentsche Land hangt voor altijd in de lucht. Daaruit jullie in petto heeft! spreekt zoveel verbondenheid! Of ik nu tweehonderd kilometer verderop woon of Ook buiten Karakter heb ik veel leuke en slimme promovendi ontmoet. niet, de lijntjes met jullie zijn ijzersterk. Jesper, Karlijn, Daan en Ties, ik vind jullie Sanne, jij maakte IMFAR in San Sebastian tot een feest! Boudewijn, Daan, Daniël, geweldig. Met geen mogelijkheid had ik meer van jullie kunnen genieten dan ik Danique, Denise, Desirée, Janna, Marloes, Marten, Melanie en Niels bedankt voor heb gedaan! de maandagmiddagmeetings en social-e-vents. Pap en mam, een dochter kan zich geen lievere ouders wensen dan jullie. Extrèmmers Andrieke, Janneke, Marloes en Vera, stelletje geweldige allrounders! Jullie zijn liefhebbende ouders pur sang. Ik heb grote bewondering voor jullie Dat ook jullie psycholoog en/of promovenda zijn maakt onze dates soms tot ongelimiteerde betrokkenheid en zorgzaamheid, en ben jullie heel dankbaar voor halve werkoverleggen. Ik geniet echter nog veel meer van ons scala aan andere het warme thuis dat jullie altijd geboden hebben en nog altijd bieden. gedeelde interesses. Wanneer ronden we een gespreksonderwerp nu eindelijk eens af, zonder dat we er ‘even tussendoor...’ nooit meer op terugkomen? Hoewel Sander, aanhoudende jeugdliefde. Op macroniveau betekent het niets, op ik werkelijk mijn bést moet doen om verbaal boven jullie uit te komen, houd ik erg microniveau betekent het alles. Alles. Ik ben je dankbaar dat je me altijd van onze groepsdynamiek. Laten we gauw weer samen op pad gaan! aangemoedigd hebt to go explore. Je hebt mijn wereld verrijkt en verruimd. Nu het promotietraject afgerond is komt er weer ruimte voor iets anders. Ik heb heel Andrieke, vele, vele uren hebben we samen in ons kleine kantoor doorgebracht. Op sommige dagen besteedden we uren aan functionerings- 272 veel zin in een nieuw avontuur samen met jou; wat zullen we gaan doen?! 273 About the author 274 Jolanda van der Meer (1983) was born in Enschede. She moved to Nijmegen in 2003 to study psychology at the Radboud University. During college, Jolanda worked as a student assistant at the Max Planck Institute for Psycholinguistics, where she coordinated and implemented a study on language acquisition in children 6 to 12 years of age. After obtaining her bachelor’s degree, she decided to aim for a master’s degree in the field of neuropsychology and rehabilitation psychology. She wrote her master thesis at Karakter University Centre for Child and Adolescent Psychiatry in Nijmegen. The focus of her thesis was on the attentional bias towards mood-congruent emotional stimuli amongst depressed and nondepressed adolescents. After thesis completion, Jolanda spent half a year in Jakarta, Indonesia. She taught neuropsychology at the Atma Jaya University and coordinated the data-collection for several cognitive tasks across the Indonesian child and adolescent population. At the same time, she volunteered for Yayasan Kampung Kids and Werkgroep ’72. Consecutively, she completed her master’s degree cum laude in 2009 with an internship at the department of Medical Psychology of the Radboud University Nijmegen Medical Centre, and continued working there. In parallel, she commenced with her Ph.D-project, which resulted in this thesis. During her Ph.D, she attended multiple courses and gave lectures, organised a symposium on the relationship between clinical practice and academic research, participated in the writing of a successful research proposal for innovative ADHD-treatment, and presented at international conferences. Noteworthy are her conference talks at the International Meeting for Autism Research (IMFAR) and the European Network of Hyperkinetic Disorders (EUNETHYDIS); for the latter she won the Sagvolden Scholarship. In her leisure time she committed herself to a local political party and voluntarily attended demented people. Currently, she works as a senior policy maker at the Dutch Knowledge Centre for Child and Adolescent Psychiatry on the changes in the Dutch child welfare system (i.e. decentralization and transformation) by 2015. In addition, she recently started her own business, Therapeutic Smile, for the treatment of ASD and ADHD affected children and adolescents. 277 Chapter 12 Jolanda van der Meer (1983) werd geboren te Enschede. In 2003 verhuisde zij naar Hier richt zij zich op de aanstaande decentralisatie en transformatie van de Nijmegen om daar psychologie te studeren aan de Radboud Universiteit. Naast kinder-en jeugdpsychiatrie. Daarnaast heeft zij onlangs als zelfstandig gevestigd haar studie was zij als student-assistent werkzaam bij het Max Planck Instituut psycholoog Therapeutic Smile opgestart, om kinderen en adolescenten met ASS voor psycholinguïstiek. Op de afdeling taalverwerving heeft zij een taaltraining en ADHD te behandelen. opgezet en uitgevoerd bij kinderen tussen de 6 en 12 jaar oud. Na het afronden van haar bachelor koos zij voor de master neuro-en revalidatiepsychologie. Haar afstudeerscriptie heeft zij geschreven bij Karakter, expertisecentrum voor complexe kinder- en jeugdpsychiatrie te Nijmegen. Haar scriptie richtte zich op de invloed van stemming op de aandacht voor emotioneel geladen stimuli bij depressieve en niet-depressieve adolescenten. Na afronding van haar scriptie heeft Jolanda een half jaar gewerkt aan de Atma Jaya Universiteit te Jakarta, Indonesië. Hier heeft zij onderwijs verzorgd in de neuropsychologie en de coördinatie gevoerd over de dataverzameling van cognitieve taken ter normering voor de Indonesische kinder-en jeugdpopulatie. Daarnaast heeft zij aldaar vrijwilligerswerk gedaan voor Yayasan Kampung Kids en Werkgroep ’72. Terug in Nederland heeft Jolanda haar klinische stage doorlopen aan het Radboud UMC te Nijmegen, op de kinderafdeling van medische psychologie. In 2009 is zij cum laude afgestudeerd, waarna zij als neuropsycholoog werkzaam bleef op de afdeling medische psychologie. Parallel hieraan is zij gestart aan het promotieonderzoek dat geresulteerd heeft in dit proefschrift. Tijdens dit traject heeft zij verscheidene cursussen gevolgd, onderwijs verzorgd, een symposium georganiseerd over de brug tussen de klinische praktijk en wetenschappelijk onderzoek, meegeschreven aan een succesvolle ZonMw subsidieaanvraag voor een innovatieve behandeling van ADHD, en gepresenteerd op internationale congressen. Hoogtepunten waren haar presentaties op de International Meeting for Autism Research (IMFAR) en European Network of Hyperkinetic Disorders (EUNETHYDIS), voor laatstgenoemde heeft zij de Sagvolden beurs gewonnen. In haar vrije tijd was Jolanda actief betrokken in de lokale politiek en deed zij vrijwilligerswerk met dementerende ouderen. Momenteel is Jolanda werkzaam als senior beleidsmedewerker bij het Kenniscentrum voor kinder-en jeugdpsychiatrie. 278 279