The relationship between parent-rated executive dysfunction and social difficulties in children and adolescents with autism spectrum disorder

Tonje Torske

Thesis for the Degree of Philosophiae Doctor (PhD) Department of Psychology, Faculty of Social Sciences, University of Oslo.

The research was conducted at the Division of Mental Health and Addiction, Vestre Viken Hospital Trust, Drammen, Norway

Asker, 2020

© Tonje Torske, 2020

Series of dissertations submitted to the Faculty of Social Sciences, University of Oslo No. 796

ISSN 1564-3991

All rights reserved. No part of this publication may be reproduced or transmitted, in any form or by any means, without permission.

Cover: Hanne Baadsgaard Utigard. Print production: Reprosentralen, University of Oslo.

Table of contents

Acknowledgements ...... 3 List of abbreviations ...... 5 Summary ...... 6 List of papers ...... 8 1. Introduction ...... 9 1.1 Autism spectrum disorder (ASD) ...... 9 1.1.1 Historical perspective ...... 9 1.1.2 Diagnosis and symptoms ...... 9 1.1.3 Epidemiology ...... 11 1.1.4 Sex differences ...... 12 1.1.5 Comorbidity ...... 12 1.1.6 Etiology ...... 13 1.1.7 Polygenic scores ...... 15 1.1.8 Neurobiology ...... 16 1.1.9 Treatment ...... 17 1.1.10 Outcome and function ...... 18 1.2 Cognition ...... 20 1.2.1 Cognitive theories of ASD ...... 21 1.2.2 Executive Function (EF) ...... 22 1.2.3 Executive dysfunction and ASD ...... 24 1.2.4 Measures of EF ...... 25 1.2.5 The relationship between EF and intelligence ...... 26 1.2.6 Comorbid ADHD and executive dysfunction ...... 27 1.3 Social function and the relationship to EF ...... 27 1.4 Unanswered questions ...... 29 2. Aims ...... 30 3. Methods ...... 31 3.1 Design ...... 31 3.2 Participants ...... 31 3.3 Procedures ...... 33 3.4 Measures ...... 34 3.4.1 Clinical assessment ...... 34 3.4.2 Questionnaires ...... 35 3.4.3 Polygenic score (PGS) ...... 37 3.4.4 Copy-number variation (CNV) ...... 38 3.5 Statistical analyses ...... 38 3.6 Ethical considerations ...... 40 4. Summary of papers ...... 41 5. Discussion ...... 44 5.1 Main findings ...... 44 5.1.1. Summary of main findings ...... 44 5.1.2 The relationship between metacognitive aspects of EF and social difficulties ...... 45 5.1.3 Sex differences in the association between EF and social difficulties ...... 49 5.1.4 Genetic relationship to EF ...... 51 5.1.5 The association to comorbid ADHD ...... 54 5.1.6 Intelligence and EF ...... 56

1 5.1.7 General considerations about the association between executive dysfunction and social difficulties in ASD ...... 56 5.2 Methodological issues ...... 58 5.2.1 Representativeness and generalizability of results ...... 58 5.2.2 Possible confounding factors ...... 59 5.2.3 Measurements ...... 60 5.2.4 Questionnaires versus neuropsychological testing ...... 63 5.3 Implications ...... 64 5.3.1 Theoretical implications ...... 64 5.3.2 Clinical implications ...... 64 5.4 Strengths and limitations, and future research ...... 67 6. Conclusion ...... 68 7. References ...... 68

2 Acknowledgements My sincere appreciation and gratitude to my main supervisor Terje Nærland, PhD, for great discussions, for giving me concrete suggestions how to move forward, and for always having something positive to say when I asked for advice. I am deeply grateful to Professor Ole A. Andreassen for sharing his extensive research experience, sharp logic, hands-on input to papers and for letting me be one of the first PhD students on the BUPGEN project. Thank you to Professor Merete G. Øie for her keen insight into neuropsychology, the field closest to my heart and core profession. Her knowledge, availability and straightforwardness have been invaluable. To my research group and co-authors: Anett Kaale, Anne Lise Høyland, Eva Malt, Jarle Johannessen, Sigrun Hope, and Thomas Bjella. Thank you for all the great discussions, the mutual research interest and the good team spirit. To Nina Stenberg and Ruth Hypher for being co-authors on research papers and for being inspirational clinicians and researchers in the field. To Anne Falsen and Karen Marte Olsen for your positive attitude and help in including patients from the Oslo University Hospital. To Liliana Buer and Thomas Bjella for being effective, positive and helpful with all issues relating to the database. To Daniel S. Quintana for assistance with statistical analyses, proofreading and co-authorship. To Francesco Bettella for support with biostatistics, interpretation of polygenic scores and critical contributions as co-author. My gratitude to Kjell Tore Hovik for help in proofreading the final document.

To Vestre Viken for letting me do research in the clinic with a patient group I truly care about and for letting me use my clinical experience in research. To Carl-Aksel Sveen, Heidi Taksrud, Kristin Sørbreda Breda, Lars Hammer, Paul Møller and Roar Fosse from Vestre Viken for making it possible for me to take a PhD. I hope that I will be able to bring research back to the clinic. To my team in the clinic Ingrid Sjaastad, Line Ulstein, Marianne Bahner, Marte Tandberg and Randi Solli for their tremendous help in including patients and for discussing clinical challenges relating to the patient group. To all the BUPGEN participants and their parents and families for their willingness to contribute to important research in the field. To the Institute of Psychology at the University of Oslo for allowing me to attend their excellent PhD program.

To my good friend and colleague Solveig L. Hauger, for being my mentor and a faithful source of mental support throughout my PhD effort. To my dear mother Gerd Torske for

3 always encouraging me to prioritize education, and for reminding me that food, sleep and balance in life is important. To my dear late father for teaching me how to strive to do my best in my field and to value hard work, and for sharing the love of math and numbers. To my sister Gøril Torske for always being a great support, for being such an inspiring person and for reminding me of all the things I love in life. To Karl Erik Olsen for taking interest in my research and for showing me that research has many commonalities across research fields, and for introducing me to PhD comics. To the late PhD Bjørg Røed Hansen for introducing me to the field of psychology early in my life and for being such an inspiration as a friend, clinician and researcher.

To my dear husband Andreas E. Hansen, for being my partner in life, for all the love and support, and for making sure we have some nice breaks. To our beloved boys Mathias, August and Jakob. Thank you for your curiosity and love of life. It makes me so happy that you are so eager to learn about the world around you, and that you come up with new and creative suggestions for how we can solve future problems. Research is formalized curiosity!

The research was funded by Vestre Viken Hospital Trust.

4 List of abbreviations ADHD = Attention deficit/ hyperactivity disorder ADI-R = The Autism Diagnostic Interview - Revised ADOS = Autism Diagnostic Observation Schedule ASD = Autism Spectrum Disorder BRI = Behavioral Regulation Index BRIEF = Behavior Rating Inventory of Executive Function CBT = Cognitive Behavioral Therapy CC = Central Coherence CNV = Copy-number variation DSM-5 = Diagnostic and Statistical Manual of Mental Disorders - Fifth Edition EF = Executive Function GEC = Global Executive Composite GWAS = Genome-wide association studies ICD-10 = International Classification of Diseases – version 10 INT = general intelligence IQ = Intelligence Quotient MI = Metacognition Index PC = Principal Component PDD-NOS = Pervasive Developmental Disorder – not otherwise specified PGS = Polygenic score SCQ = Social Communication Questionnaire SD = Standard Deviation SNPs = Single-Nucleotide Polymorphisms SRS = Social Responsiveness Scale TD = Typically Developed ToM =Theory of Mind

5

Summary The relationship between executive function (EF) and social difficulties in children and adolescents with autism spectrum disorder (ASD) is investigated in the current thesis. Executive dysfunction is one of the cognitive theories of ASD. Even though it cannot explain all of the symptoms of ASD, executive dysfunction is a central characteristic of everyday functioning in ASD and closely related to adaptive functioning and quality of life. Few studies, however, have investigated the relationship between EF and social function in everyday life in this group. Sub-aims of this thesis were to examine how sex and comorbid attention deficit/ hyperactivity disorder (ADHD) affect the relationship between social function and EF. The role of EF by means of genetic information is also examined. We had access to polygenic scores (PGSs) for ASD, ADHD and general intelligence, and explored their association to executive dysfunction in everyday life.

The first aim was to investigate the relationship between everyday EF as measured by the Behavior Rating Inventory of Executive Function (BRIEF) and social function as measured by the Social Responsiveness Scale questionnaire (SRS). In a sample of 86 children and adolescents (23 girls), we found that difficulties with the metacognitive aspects of EF were significantly related to social dysfunction (Paper I). This finding remained even after the children with comorbid ADHD were excluded from the analyses. Secondly, we wanted to examine possible sex differences in the relationship between executive dysfunction as measured by the BRIEF and autistic symptoms using the Autism Diagnostic Interview- Revised (ADI-R). The results showed a strong and significant association between BRIEF scores and ADI-R domains for social interaction and communication in girls, but not for boys. We did not find sex differences in the relationship between executive dysfunction and restricted and repetitive behaviors (Paper II). Comorbid ADHD did not have a significant impact on the association between BRIEF and ADI-R scores.

The last aim was to explore if the polygenic score (PGS) for ASD was associated with executive dysfunction in everyday life in a clinical sample of children and adolescents admitted for clinical assessment of ASD. In addition, we had access to the PGSs for ADHD and general intelligence (INT) for the same participants. The sample was divided into low and high groups based on their PGSs. In a regression model, we found that ASD PGS was significantly associated with behavior regulation aspects of EF. We did not find any

6 significant association between EF and the PGSs for ADHD or INT in our sample (Paper III). Furthermore, we found high PGS for general intelligence to be related to social difficulties in everyday life measured by the SRS.

One possible clinical implication of our findings is that metacognitive aspects of EF may be of particular importance for social function in everyday life for children and adolescents with ASD. Interventions targeting metacognitive skills could therefore have a positive influence on social function for these children.

Furthermore, there might be important differences between girls and boys in the relationship between executive dysfunction and reciprocal social interaction and communication. Our results indicated that the relationship for girls was stronger than for boys. This might imply that executive dysfunction and social difficulties in girls may be more closely related, and that girls with ASD may benefit more from EF interventions.

Lastly, children and adolescents with high PGS for ASD might be particularly vulnerable and have more difficulties with behavior regulation aspects of EF than those with low PGS for ASD. The clinical relevance may be that preventions aimed at EF difficulties could be offered to children at risk earlier, and that the more general ASD treatment could be stratified according to PGS level.

7 List of papers The thesis is based on the following papers, referred to in the text by their Roman numbers I- III.

Paper I Torske, T, Nærland, T., Øie, M. G., Stenberg, N. & Andreassen, O.A. Metacognitive aspects of executive function are highly associated with social functioning on parent-rated measures in children with autism spectrum disorder. Published in Frontiers in Behavioral Neuroscience, 10 January 2018 .

Paper II Torske, T., Nærland, T., Hypher, R.E., Kaale, A., Høyland, A-L., Hope, S., Johannessen, J., Øie, M.G. & Andreassen, O.A. Sex differences in the relationship between social difficulties and executive dysfunction in children and adolescents with autism spectrum disorder. Under review in Scientific Reports. Preprint published on BioRxiv the 20th of December 2018: https://doi.org/10.1101/501932.

Paper III Torske, T., Nærland, T., Bettella, F., Bjella, T., Malt, E., Høyland, A.L., Stenberg, N., Bjella, T., Øie, M.G., & Andreassen, O.A. Autism spectrum disorder polygenic scores are associated with every day executive function in children admitted for clinical assessment. Published in Autism Research 30 September 2019 https://doi.org/10.1002/aur.2207.

8

1. Introduction 1.1 Autism spectrum disorder (ASD) 1.1.1 Historical perspective The writings of Kanner and Asperger are considered to be the first descriptions of autism and autistic behavior. In 1943, Leo Kanner wrote his famous paper entitled “Autistic Disturbances of Effective Contact” where he describes the behavior of 11 children (including three girls) and referred to it as a distinct syndrome (Kanner, 1943). The behavior he described was a pattern of social disinterest and self-isolation in combination with an insistence on sameness and resistance to change. The term autism was borrowed from the Swiss psychiatrist Eugen Bleuler who had used it to describe self-centered thinking in schizophrenia (Bleuler, 1911). However, while Bleuler used the term to describe excessive hallucinations and fantasy in infants, child psychologists started using the term autism in the 1960s to refer to children with a lack of fantasy and unconscious symbolic life (Evans, 2013). The year after Kanners paper, Hans Asperger published a German paper about four children whom he thought had “autistic psychopathology” (Asperger, 1944). The four boys described by Asperger were reported as having social difficulties combined with in-depth knowledge in specific areas, and he called them “young professors”. This later developed into the description of Asperger syndrome, where people have intact intellectual abilities, but distinct problems with reciprocal social interaction and restricted patterns of behaviors and interests. Later, Lorna Wing described autistic symptomologies as “a spectrum” (Masi, DeMayo, Glozier, & Guastella, 2017; Wing, 1981). Furthermore, she defined the triad of autism that consisted of dysfunction within three areas: social, communicative and restricted behaviors/interests. She also described how these children lacked the ability to pretend play. In the third version of the Diagnostic and Statistical Manual of Mental Disorders (DSM-III) published in 1980, “autism” became a diagnostic entity in its own right and the condition was grouped under the label Pervasive Developmental Disorders (PDD) (Tidmarsh & Volkmar, 2003).

1.1.2 Diagnosis and symptoms ASD is currently defined in terms of behavior. ASD is characterized by persistent deficits in social interaction and social communication across multiple contexts, and a restricted, repetitive pattern of behaviors, interests, or activities (American Psychiatric Association, 2013). In the newest diagnostic system, Diagnostic and Statistical Manual of Mental Disorders - Fifth Edition (DSM-5), the triad becomes a dyad where difficulties with social

9 interaction and social communication are collapsed together, and stereotype behaviors and interests make up the other part of the dyad (American Psychiatric Association, 2013). Furthermore, the term Asperger and other subcategories are removed, and they are all grouped under the term ASD. In the diagnostic manual DSM-5, ASD has become a spectrum diagnosis that can range from very mild to severe, and more people with the condition have normal intelligence and are able to speak, read and live in the community (Lord, Elsabbagh, Baird, & Veenstra-Vanderweele, 2018). However, the core symptoms of ASD have not changed substantially from the original description.

The diagnostic manual The International Classification of Diseases – version 10 (ICD-10) describes the same characteristics, with difficulties relating to social communication and social interaction across multiple contexts, and a restricted, repetitive pattern of behaviors, interests, or activities, as in DSM-5 (World Health Organization, 1992). These difficulties are defined as qualitative abnormalities of behavior that are deviant in relation to mental age and are of a pervasive nature that will affect the individual’s function in all situations. The ICD-10 divides ASD into the subcategories childhood autism, atypical autism, Asperger syndrome and Pervasive Developmental Disorder – not otherwise specified (PDD-NOS) (World Health Organization, 1992). In this thesis, I will use the term ASD to refer to all the subcategories that are included in ICD-10. Most people that have an ASD diagnosis based on the ICD-10 criteria will also fulfill the criteria for ASD in DSM-5. However, the diagnostic criteria in DSM-5 is somewhat stricter than ICD-10 and DSM-IV, and some people with PDD-NOS given in accordance with the ICD-10 criteria might not fulfill the requirements for an ASD diagnosis based on the DSM-5 criteria (Doernberg & Hollander, 2016). In preparations for the upcoming ICD-11 manual, the description of autism includes two categories (as in DSM-5), namely difficulties in social interaction and communication, and restricted and repetitive behaviors and interests.

ASD is defined as a neurodevelopmental disorder, meaning that the disorder manifests in childhood in the developing brain (American Psychiatric Association, 2013). However, it is not restricted to childhood and adolescence. For most people with ASD, it is a lifelong condition that can have different expressions and challenges during different developmental phases (Lord et al., 2018).

10 1.1.3 Epidemiology Prevalence is defined as the proportion of individuals in a population who suffer from a defined disorder at any given time, and the global prevalence of ASD in 2012 was estimated by the World Health Organization to be 1% (Elsabbagh et al., 2012). However, a recent review estimated the prevalence to be 1.5% in developed countries (Lyall et al., 2017). A recent study confirmed that the prevalence of ASD in school-aged children in China is about 1%, which is comparable to western prevalence estimates (Sun et al., 2019). In the United States, an association has been found between higher socio-economic status and higher ASD prevalence, and especially ASD without intellectual disability. Some of the explanation for this has been suggested to be related to better access to health care systems and diagnostic services among those with higher socio-economic status; no plausible biological mechanism has been found for this difference in prevalence. In other countries with more equal access to health services, ASD has been associated with lower socio-economic status (Myers, Chavez, Hill, Zuckerman, & Fombonne, 2019). Some have found that racial/ethnic differences affect diagnostic rates above and beyond the effect of socio-economic status. However, this phenomenon is mostly found in US-based populations. For example, Somali children are more often diagnosed with ASD than other ethnic groups in both the USA and Sweden, and they are also more often found to have comorbid intellectual disability. One possible causal explanation for this has been suggested to be lower levels of vitamin D among immigrant Somali mothers that can impact fetal brain development (Myers et al., 2019).

The prevalence rate has been rising considerably since the 1970s, and the most recent data from the Centers for Disease Control and Prevention show that 1 in 59 children in the United States have ASD (Baio, Wiggins, & Christensen, 2018). Several studies have documented an increase in incidence of ASD, which is defined as the probability of new cases of a disorder occurring within a given period of time (Davidovitch, Hemo, Manning-Courtney, & Fombonne, 2013). In the United Kingdom, the annual incidence rate is approximately 1 per 1000 (Jick, Beach, & Kaye, 2006). There may be several reasons for this increase. Researchers point to better detection of behavior patterns associated with ASD and a broadening of the diagnosis criteria as the most important reasons. The use of different samples and statistics makes it hard to compare different surveys and prevalence numbers (Fombonne, 2018).

11 1.1.4 Sex differences Increasing attention has been paid to understanding possible gender and sex differences in ASD (Halladay et al., 2015; Lord et al., 2018). One of the most consistent findings is that ASD is overrepresented in boys compared to girls. Traditionally, the male-to-female ratio is thought to be 4:1 (Lai, Lombardo, Auyeung, Chakrabarti, & Baron-Cohen, 2015). The reasons for this, however, are not fully understood. A recent meta-analysis of population- based ASD studies concluded that the male-to-female ratio is closer to 3:1 (Loomes, Hull, & Mandy, 2017). The prevalence ratio is different depending on intellectual level, where a lower intelligence quotient (IQ) is associated with a lower male-to-female ratio. Several have suggested that girls might have a different phenotype of ASD than boys. Furthermore, the diagnostic instruments have been primarily developed to capture the typical male expression of ASD. It has also been argued that girls camouflage their autistic symptoms, making is harder to recognize autistic symptoms in girls (Lai et al., 2017). Others hold a different view. They argue that it is not deliberate camouflaging, but rather that girls use compensatory behaviors that make their autistic behaviors less apparent (Dean, Harwood, & Kasari, 2017). Furthermore, girls referred to the clinic for assessment of ASD often have more cognitive and behavioral problems than boys. Dworzynski et al. argue that it is more difficult for clinicians to detect average IQ girls without additional behavior problems and more subtle forms of ASD than boys (Dworzynski, Ronald, Bolton, & Happe, 2012). This might be due to the fact that it seems that girls with ASD have an increased mutational burden, which indicates an elevated threshold for developing ASD in girls (Jacquemont et al., 2014). This has been interpreted as a female protective effect, in other words, a greater resistance to ASD from genetic causes in females (Levy et al., 2011). A possible consequence of an increased genetic load in girls is that those who reach a clinical diagnosis of ASD often have lower intelligence and more behavioral problems than boys with ASD (Dworzynski et al., 2012).

1.1.5 Comorbidity ASD often co-occurs with other developmental delays, such as intellectual disability, communication disorders, specific learning disorders, and motor disorders (Gillberg, 2010). The latest diagnostic manual (DSM-5) takes account of the fact that ASD often co-occurs with other psychiatric conditions. Attentions deficit/ hyperactivity disorder (ADHD) is the most common comorbidity in people with ASD, and about 30% with ASD have a comorbid diagnosis of ADHD (Lord et al., 2018). As ASD, ADHD is also a neurodevelopmental disorder and is more common in boys than in girls (Thapar & Cooper, 2016). Furthermore,

12 having a comorbid diagnosis of ADHD negatively affects the outcome of children with ASD and may be a target for interventions. Anxiety and depression are also common in ASD and are reported to increase in adolescence, especially for girls (Lord et al., 2018). Tic disorders, obsessive-compulsive disorder and eating disorders are common psychiatric comorbidities in children with ASD (Simonoff et al., 2008).

In 2010, Gillberg coined the conceptual framework called ESSENCE to emphasize co- existence and comorbidities of neurodevelopmental disorders. ESSENCE is an acronym for Early Symptomatic Syndromes Eliciting Neurodevelopmental Clinical Examinations (Gillberg, 2010). By introducing this term, he wanted to emphasize that the existence of comorbidities is the rule rather than the exception in child and adolescent psychiatry, and to highlight the importance of seeing the conditions in relation to each other and not just as separate diagnoses (Gillberg, 2010).

1.1.6 Etiology Environmental risk factors The underlying cause of developing ASD is most likely an interaction between genetic liability and environmental factors. Among the environmental factors during pregnancy commonly linked to ASD are the following: malnutrition, infections, medications, pollutants, higher parental age, parental obesity, preterm birth and low gestational age (Lord et al., 2018). A recent review found evidence for the maternal factors of advanced age (≥35 years), chronic hypertension, pre-eclampsia, gestational hypertension, and overweight before or during pregnancy to be convincingly associated with ASD (Kim et al., 2019). On the other hand, a meta-analyses confirmed that maternal use of folic acid supplements during pregnancy significantly reduced the risk of ASD in children regardless of ethnicity (Wang, Li, Zhao, & Li, 2017).

The “refrigerator mother theory” (emotionally cold mothers) is discarded as a cause of ASD, and there is no evidence that a lack of “genuine maternal warmth” can make children autistic (Bettelheim, 1967; Kanner, 1949). Furthermore, the idea that the vaccine for measles, mumps and rubella (MMR) causes ASD has been thoroughly refuted. Wakefield and colleagues published a paper in the Lancet in 1998 suggesting that the MMR vaccine could lead to ASD. This paper was later retracted due to breaches of ethical standards and scientific misrepresentation. Wakefield was found guilty of deliberate fraud (Rao & Andrade, 2011).

13

Genetic factors The heritability of ASD is high, with estimates ranging from 70-90%, and ASD is regarded as a complex genetic disorder with multifactorial causes (Lai, Lombardo, & Baron-Cohen, 2014). A recent multinational population study based on over 2 million participants found heritability estimates for ASD to be approximately 80% (Bai et al., 2019). Michael Rutter, a leading child psychiatrist from the UK, was the first to conduct a genetic study of ASD (Evans, 2013). Twin studies were the first studies to prove the genetic role in the development of ASD. Now knowledge has been gained from both genetic epidemiological (e.g. family and twin studies) and molecular genetic research projects (e.g. genome wide association studies (GWAS)). However, while the genetic contribution to the disease (heritability) is high for ASD, the pathophysiology is still mostly unknown (Editorial, 2010; Manolio et al., 2009). Moreover, the understanding of the gene-environment interplay is still at an early stage (Lai et al., 2014). Though many genetic risk loci have been identified, they only explain a fraction of disease liability (Bourgeron, 2016; Lai et al., 2014). Linking genetic vulnerabilities to clinical symptoms has the potential to provide a better understanding of the underpinnings of the disorder.

ASD is related to a range of known medical genetic conditions and syndromes like Fragile X syndrome, Cornelia de Lange syndrome, Tuberous Sclerosis Complex, Neurofibromatosis type 1, Downs syndrome, and DiGeorge syndrome/ 22q11. Copy-number variations (CNVs) have in molecular studies been identified as risk factors for ASD (Geschwind, 2011). While several rare genetic variants conferring a high risk of ASD have been identified, they only explain about 5-10% of the overall genetic risk factor (de la Torre-Ubieta, Won, Stein, & Geschwind, 2016; Ramaswami & Geschwind, 2018). Most of the heritability of ASD seems to be caused by common genetic variants (Gaugler et al., 2014).

Genetic factors underlying the skewed sex ratio in ASD remains mostly unknown, and cannot be explained by x-linked variants, since most known ASD risk genes are located in autosomal regions (Werling & Geschwind, 2013). There is some support for a female protective effect, which means that females have a higher liability threshold for ASD than men. However, the biological mechanism behind the female protective effect is not yet well understood. Furthermore, females have a higher rate of de novo risk variants than males, which cannot

14 solely be explained by a greater frequency of females with lower IQ in these samples (Werling, 2016).

In summary, genetic evidence indicates that ASD is not a disorder with a single causal mechanism, but a number of etiologically distinct conditions with diverse pathophysiological mechanisms leading to similar behavioral manifestations (de la Torre-Ubieta et al., 2016). Furthermore, it might also be a case of pleiotropy, where one gene influences two or more seemingly unrelated disorders (Rommelse, Geurts, Franke, Buitelaar, & Hartman, 2011). In this way, the same gene defect can give rise to ASD in one person and ADHD or schizophrenia in another person (Geschwind, 2011).

1.1.7 Polygenic scores To better capture the polygenic nature of complex disorders, polygenic score (PGS) can be calculated from large genome-wide association studies (GWAS) and used to estimate the overall genetic risk of a given phenotype in a given genotyped individual. PGS is a statistically computed estimate of the cumulative genetic risk in an individual, and has recently been introduced as a tool to explore the associations between genes, symptoms and functioning in mental illnesses such as schizophrenia, bipolar disorder, ADHD and ASD (Fanous et al., 2012; International Schizophrenia Consortium et al., 2009). Summary scores are based on GWAS discovery samples where millions of single-nucleotide polymorphisms (SNPs) have been scanned in order to identify those alleles that distinguish cases from controls in the particular phenotype or disorder being studied. This information is then used to generate the PGS from the GWAS phenotypes, which can be used to calculate the PGS of each individual in an independent “target” sample (Palk, Dalvie, de Vries, Martin, & Stein, 2019). Currently, PGSs have been calculated for several disorders and traits that can be of particular interest to the ASD population, such as the PGS for ASD, ADHD, general intelligence and depression (Demontis et al., 2018; Grove et al., 2019; Savage et al., 2018; Wray et al., 2018). The samples related to ASD are now large enough for the first common variants to be detected (Grove et al., 2019). In the future, it is expected that several SNPs, which individually have small effects, will be detected. Overall, these will help explain more of the heritability of ASD. In this thesis, I will be using the term polygenic score (PGS) instead of polygenic risk score, because the traits/conditions we are studying are not always a disadvantage. It is not appropriate to talk about “the risk” of an intelligence or height score.

15 Even though the sensitivity and specificity of the ASD PGS is not yet high enough for clinical use in diagnostic and treatment planning in ASD, it is thought to have some clinical potential in the future (Grove et al., 2019). In somatic diseases like coronary artery disease, for example, PGS has been used -- independent of family history -- to identify high-risk individuals who benefit from medical treatment with statins. PGSs have also shown utility in decisions relating to disease screening for breast and prostate cancer (Torkamani, Wineinger, & Topol, 2018). Exploring the clinical use of PGS is central in the pursuit of precision medicine for neurodevelopmental disorders and could be used as one of several factors to adapt treatment to individuals. Identifying those with particularly low or high risk for a given disorder may be a viable approach. People with a high risk of developing ASD might also have an increased risk of developing known comorbid disorders, and the combination of high risk for both ASD and ADHD may be important in guiding interventions. If children at risk can be identified early, it might be of clinical relevance to initiate prevention interventions aimed at specific difficulties, or stratify the more general ASD treatments by PGS. Intervention studies are needed to investigate responders versus non-responders based on PGS in combination with clinical information.

1.1.8 Neurobiology The neurobiological mechanisms of ASD are not yet fully understood. Early brain overgrowth, particularly under the age of 6 years, is the most consistently reported neuroanatomical finding in ASD (Parellada et al., 2014; Stanfield et al., 2008). Compared to typically developed (TD) controls, people with ASD have often been reported to have increased volume in the frontal and temporal lobes, increased cortical thickness in the frontal lobe, increased surface area and cortical gyrification, increased cerebrospinal fluid volume, as well as reduced cerebellum volume and reduced corpus callosum volume (Pagnozzi, Conti, Calderoni, Fripp, & Rose, 2018).

ASD is also associated with altered brain connectivity. Most findings show a pattern of overall brain under-connectivity, combined with over-connectivity particularly in frontal and occipital regions (Lord et al., 2018). However, the underlying cellular mechanisms for these patterns remains unclear. Different sensitivity to the environment and distinct learning styles among people with ASD might also lead to brain reorganization during development. Since ASD has multifactorial causes and comorbidity is common, it is likely that brain locations and neural circuitries are affected differently from person to person. These neurobiological

16 changes are likely to be age-dependent and subject to developmental trajectories (Gillberg et al., 2019; Lai et al., 2014).

There is also some evidence of sex differences in ASD on a neurobiological level. Lai et al. report differences in neuroanatomy between females and males with ASD. They suggest that females with ASD, but not males, show a neuroanatomical “masculinization”. This means that females with ASD have neuroanatomical abnormalities that overlap with areas showing typical sexual dimorphism in controls (Lai et al., 2013). The difference in neuroanatomy is not found in males with ASD compared to TD males. How these differences in neuroanatomy are associated with cognition is still unclear.

1.1.9 Treatment Beneficial treatments for ASD with the most empirical evidence are behavioral, developmental and/or educational, and these types of interventions are considered treatment of choice for children and youth with ASD (Odom, Morin, Savage, & Tomaszewski, 2019). Early parent-mediated treatments have shown some effectiveness, and the effect size for these types of interventions are approximately d = 0.30 (a small effect size). Early naturalistic developmental behavioral interventions, such as Applied Behavioral Analysis, have received most attention. The treatments are usually intensive (15-20 hours per week) and emphasize play, social interaction and communication initiation. A meta-analysis of treatment studies of these kinds of approaches reported effect sizes of d = 0.69 for adaptive skills, d = 0.76 for IQ and about d = 0.50 for language skills after 2 years of treatment (Lord et al., 2018). For school-aged children and adolescents, social skills groups are the most common intervention, and here there are many different types of approaches and manual training programs. Randomized controlled trials have documented substantial self-reported improvements in social skills and tasks using these kinds of manual-based programs (d = 0.98 and d = 0.58, respectively), yet lower effect sizes when looking at parent and teacher reports (Lord et al., 2018). The programs are often based on cognitive behavioral therapy techniques, and some target EF deficits in particular (Kenworthy et al., 2014).

A known challenge is that most children receive their treatment in community-based services, and that most evidence-based treatments are studied and given in university and/or hospital clinics. A recent meta-analysis reported a large gap between treatment outcomes in

17 community-based services and those reported in efficacy trials (Nahmias, Pellecchia, Stahmer, & Mandell, 2019).

While no medical treatments have proven effective for core social and communicative impairments in ASD, some medications can improve behavior problems and co-occurring conditions. Irritability, physical aggression, self-injurious behavior and severe tantrums can be reduced with drugs like risperidone and aripiprazole. Furthermore, it is possible to use medication for comorbid conditions, like methylphenidate for ADHD (Mooney, Fosdick, & Erickson, 2019). However, it is worth noting that the effect is often smaller and side-effects often larger for those with ASD than for children in the general ADHD population (Lord et al., 2018). Furthermore, there have been some positive results in studies using oxytocin to improve social skills (Cai, Feng, & Yap, 2018; Quintana et al., 2017).

There might be differences in the effect of treatment for different subgroups, and research is sparse on whether girls/females and boys/males respond differently to treatment. Girls may require different treatment than boys in order to improve. Research into what characterizes different subgroups can be an important precursor to understanding the specific treatment needs of specific subgroups.

1.1.10 Outcome and function Even though there are now more people with ASD without intellectual disabilities and fewer that live in institutions than earlier, ASD is still a lifelong and severe condition for the majority of children with ASD, involving a heavy burden for the individual, their family and society (Bal, Kim, Fok, & Lord, 2018). The need for health care, special education and supported living is considerable throughout the lifespan of a person with ASD (Lai et al., 2014). An economic analysis estimates that ASD costs society more than any other medical condition, including cancer, stroke and heart disease (Buesher, Cidav, Knapp, & Mandell, 2014). People with ASD, and especially those with higher cognitive abilities, experience a profound discrepancy between level of cognitive ability and adaptive functioning/everyday life skills (Tillmann et al., 2019). Even for adults with ASD without intellectual disabilities, few live independently and have jobs. Educational attainment has improved for people with ASD over the past 20 years. However, there is still a discrepancy between education and employment in this group (Lord et al., 2018). In a Swedish 20-year follow-up study of 50 males with Asperger, 41% had jobs/studies, 51% lived independently and 33% reported

18 having two or more friends. Academic success was positively correlated with IQ (Helles, Gillberg, Gillberg, & Billstedt, 2017). On the other hand, Baron-Cohen and others have argued that ASD is not always associated with disability, but may also be associated with talent in non-social fields (Baron-Cohen, 2000; Happe, 1999). As many people with ASD are factual and highly rule-governed, they can be skilled in fields like physics, engineering, computer science and mathematics, and are overrepresented in technical occupations. Others take this further and argue that by only focusing on dysfunction and impairment, atypical abilities of people with ASD might not be detected. The neurodiversity movement challenges the medical models that view ASD as pathological and a disorder that should be cured (Kapp, Gillespie-Lynch, Sherman, & Hutman, 2013). They suggest that conditions like ASD are simply less common neurological presentations of natural human variation (Jackson & Volkmar, 2019). Some have recommended that it is possible to address some of the concerns made by the neurodiversity movement by switching the term “disorder” with the term “disability” (Baron-Cohen, 2017). In this way, the focus will be less on dysfunction, since the term disability is more contextually dependent. To what extent it will be viewed as a disability will depend on the social acceptance of the mental condition in society.

The focus in ASD research has mainly been on children, adolescents and young adults, and we still know relatively little about adulthood and aging for individuals with ASD. Most follow-up studies confirm that a majority of adults with ASD continue to experience significant challenges throughout their lives (Magiati & Howlin, 2019). Increased mortality in ASD has been observed for a number of different causes with an odds ratio of approximately 2.5 compared to the general population. Causes of death are associated with several mental and behavioral disorders, diseases of the nervous, circulatory, respiratory and digestive systems, as well as congenital malformations. The risk for premature mortality in females with low-functioning ASD is particularly high, and individuals with high-functioning ASD have a higher risk of suicide (Hirvikoski et al., 2016).

Since ASD is a neurodevelopmental disorder, it is important to get access to help and treatment early to facilitate optimization of opportunities for social interaction, communication and education. All of these factors have an impact on relationships with friends and family and participation in school and leisure activities. Most clinicians work in accordance with the biopsychosocial model, originally proposed by Georg L. Engels. This model takes an holistic approach to understanding disease (Engel, 1977), and is thought to be

19 less reductionistic than the biomedical model. The biopsychosocial model places more emphasis on, for example, familial, cultural and socioeconomic factors, while also accounting for biological and psychological factors. The World Health Organisation adopted this model as a basis for the International Classification of Function in 2002 (Allan, Campbell, Guptill, Stephenson, & Campbell, 2006). In line with this approach, not only diagnosis and treatment, but also acceptance and inclusion in society, are very important factors needed to ensure quality of life for persons with ASD (Kapp, 2018).

1.2 Cognition Cognition is defined as “the mental action or process of acquiring knowledge and understanding through thought, experience, and the senses” (Oxford Dictionary, 2020). Cognitive processes use existing knowledge to generate new knowledge. The term cognition includes many aspects of intellectual functioning and processes such as attention, memory, working memory, perception, language, construction, reasoning, problem solving and decision making. (Lezak, Howieson, & Loring, 2004). Furthermore, there is an important distinction between general cognitive function, usually referred to as intelligence, and more specific functions like attention, perception, and memory. Lezak et al. describe that cognitive abilities are usually not directly observed, but instead inferred from behavior. These cognitive abilities are often measured using intelligence tests and specific neuropsychological tests for investigating certain mental functions. Other mental activities such as level of consciousness and activity rate/ speed of processing are related to and influence the efficiency of cognitive processes, but do not necessarily have a unique behavioral end product (Lezak et al., 2004).

Heterogeneity is the rule rather than the exception in ASD, and this makes is difficult to describe the “typical” cognitive profile of ASD (Lord, 2019). General intellectual level is of great importance for treatment choice and prognosis in people with ASD. About 45% of people diagnosed with ASD have an intellectual disability (Lai et al., 2014), and 11-65% of school-aged children with ASD are reported to have intellectual disabilities (Lord et al., 2018). On the other hand, ASD is also associated with higher than average IQ (Clarke et al., 2016).

Language deficits are common in ASD and cannot solely be explained by general IQ (Paul & Simmons, 2019). A large proportion also have EF deficits (e.g. difficulties with mental flexibility and planning), which are areas particularly important for everyday life functions

20 (Demetriou et al., 2017; Gilotty, Kenworthy, Sirian, Black, & Wagner, 2002). Therefore, knowledge about cognition, its behavioral expression and its relation to different subgroups of ASD is of importance for finding effective treatments for ASD.

1.2.1 Cognitive theories of ASD Three central cognitive theories seek to explain the core behavioral difficulties in ASD. These theories are the Theory of Mind Hypothesis, the Weak Central Coherence Theory and the Executive Dysfunction Hypothesis (Pellicano, 2011). These theories are often referred to as “single deficit” explanatory models, as they try to explain ASD with a single underlying cognitive deficit. However, none of the theories can truly explain all of the symptoms exhibited by any particular individual with ASD.

The first major cognitive theory of ASD was the description of difficulties with mentalizing, attributing mental states to oneself and others, and predicting and interpreting the behavior of others (Dennett, 1978). These abilities form the bases of the Theory of Mind (ToM) Hypothesis. In 1985, Baron-Cohen and colleagues conducted an important study in this regard where they found that 80% of their sample of children with ASD had difficulties with a “false-belief task” (Baron-Cohen, Leslie, & Frith, 1985). In 1989, Uta Frith wanted not only to focus on the challenges people with ASD had, but also incorporate their special abilities into cognitive theories of autism. She therefore formulated the Weak Central Coherence (CC) Theory, which states that people with ASD have a preference for processing local elements over the global whole, and thereby having difficulties with contexts and seeing “the bigger picture” (Frith, 1989). This often manifests itself as an “insistence of sameness”. The third theory is the Executive Dysfunction Hypothesis that was described by Pennington and Ozonoff in 1996. This theory emphasizes impairments related to planning and mental flexibility in ASD, and how this is associated to restricted and repetitive behaviors and interests in ASD (Hill, 2004; Pennington & Ozonoff, 1996).

Evidence suggests that many children with ASD have difficulties in all three areas (ToM, CC and Executive Function (EF)), and that none of them can explain all symptoms of ASD. Furthermore, none of these three areas are necessary or sufficient to be diagnosed with ASD. There is also research showing that the various difficulties affect each other during childhood development. Pellicano found evidence that EF and CC skills were predictive of children’s ToM performance, while no significant predictive relationship was found in the opposite

21 direction. They did not find developmental links between EF and CC (Pellicano, 2010). See Pellicano for a thorough review of these three cognitive theories of ASD, and an evaluation of their strengths, weaknesses and explanatory power (Pellicano, 2011).

Studying EF deficits is particularly interesting because it is predictive of other difficulties, closely associated with adaptive functioning and a possible endophenotype for several psychiatric disorders (Benca et al., 2017; Snyder, Miyake, & Hankin, 2015). Furthermore, EF deficits involve one of the three areas of cognitive impairment most closely linked to specific brain areas in the frontal region. The frontal region has been the subject of considerable research linking brain and behavior. Since the focus of this PhD thesis is EF and the Executive Dysfunction Hypothesis of ASD, the definition of EF and its central role in understanding difficulties associated with ASD will be described in more detail.

1.2.2 Executive Function (EF) EF enables the individual to disengage from the present context and to regulate thought and actions to effectuate future goal-directed behavior (Miyake et al., 2000). EF comprises several components including inhibition, working memory, flexibility, emotional control, initiation, planning, organization, and self-control (Hill, 2004; Miyake et al., 2000). The most prominent EF framework in research on neurodevelopmental disorders is Miyake and Friedman’s unity and diversity model (Miyake et al., 2000). This model describes EF as having some common components that make up the unity part of the definition, and others that are defined by more specific functions. The unity/diversity model focuses on three aspects of EF: common EF/inhibition, updating specific/working memory and shifting (See Figur 1). The inhibition factor is fully accounted for by common EF, while updating and shifting have specific components (Snyder et al., 2015). The Miyake definition was originally a model for adults, but it has also been suggested as a theoretical basis for research on EF in children (Garon, Bryson, & Smith, 2008). There is, however, more evidence for both unity and diversity in EF among adult samples, and a greater unidimensionality of EF in child and adolescent samples. For example, Karr et al. found in their review of EF that the shifting part of the definition emerges during school-age to adolescent years (Karr et al., 2018).

Properly functioning EF depends on critical neural substrates in the prefrontal regions of the brain. In particular, the dorsolateral frontal cortex has been linked to EF (Alvarez & Emory, 2006; Smolker, Friedman, Hewitt, & Banich, 2018). Although there is some disagreement

22 about the location of EF, Alvarez et. al. emphasizes the importance of not just investigating brain structure, but the importance of measuring observable behaviors that have real-world significance (Alvarez & Emory, 2006). Some researchers also make a conceptual distinction between hot and cold EF, where cold EF involves cognitive processes with little emotional salience in relation to neutral stimuli. Traditionally, neuropsychological tests have assessed cold EF processes such as inhibition, working memory, planning and mental flexibility. In contrast, hot EF tasks involve motivational and emotional aspects of cognition, e.g. delayed discounting and affective decision-making. Hot EF functions activate emotion regulation and reward systems in the orbito-frontal cortex, ventral striatum and limbic systems (Kouklari, Tsermentseli, & Monks, 2018; Zimmerman, Ownsworth, O'Donovan, Roberts, & Gullo, 2016). The most common tests used to measure hot EF are gambling tasks, the Marshmallow Test or delay discounting tasks (Zelazo & Carlson, 2012).

Figure 1. Unity and Diversity Model of Executive Function

Latent variable model where all individual tasks are combined to form latent factors. Numbers on arrows are standardized factors loadings (range 1 to -1). All nine tasks load onto a common EF factor (unity), and updating and shifting tasks also load onto their respective specific factors (diversity). Note that there is no inhibition-specific factor (i.e., inhibition task variance is fully accounted for by common EF).

Adapted from Snyder (Snyder et al., 2015).

EF involves cognitive functions that take the longest time to develop of all cognitive processes. They often continue to mature up until an individual’s late twenties, and this maturation involves myelination of the neurons in the prefrontal region of the brain (De Luca & Leventer, 2008). There is some indication that EF skills mature earlier in girls than boys, for example, that girls are less impulsive than boys when they are 8-10 years of age.

23 However, a recent review based on both human studies and animal research concluded that there is limited support for substantial sex differences in EF (Grissom & Reyes, 2019). They argue that the variation within the sexes is greater than the variation between them. An explanation for differences in EF between males and females might be that they use different strategies when solving EF tasks, and that different developmental trajectories in boys and girls influence EF performance (Grissom & Reyes, 2019).

EF deficits are associated with many psychiatric and developmental disorders, and problems relating to EF impairment can have more impact on everyday functioning than symptoms relating to the disorder itself (Benca et al., 2017). Snyder et al. argue that particularly impaired common EF/inhibition is associated with psychopathology. However, more specific factors might also have important implications for understanding EF deficits associated with psychopathology (Snyder et al., 2015). Difficulties with EF may cause problems regulating emotions, mastering school, functioning socially, following up treatment and reaching goals (Diamond, 2013). Having difficulties with working memory in everyday life can make it hard to remember and process the meaning of a long sentence or rehearsing a phone number. People with problems related to inhibition can be easily distracted and impulsive. Furthermore, people with problems related to set-shifting/mental flexibility can have difficulties with multitasking and may appear rigid in their thinking. Altogether, EF is important for goal-oriented behavior and being able to adapt to an everchanging world. Since EF is so important for everyday functioning and quality of life, and because EF deficits are associated with many psychiatric disorders, it is important to conduct research on EF and investigate how these deficits are related to psychiatric disorders and their symptoms. Furthermore, research indicates that it is possible to improve EF, making it a promising target for interventions (Diamond, 2013).

1.2.3 Executive dysfunction and ASD One of the major challenges for the Executive Dysfunction Hypothesis of ASD is that EF impairment is not unique to ASD, but also common in many other developmental and psychiatric disorders (Pellicano, 2011). Traditionally, it has been assumed that children with ASD have difficulties with mental flexibility and verbal fluency, but that they perform within the normal range on inhibition tests. Recent findings challenge this assumption, reporting inhibitory deficits in individuals with ASD as well (Geurts, de Vries, & van den Bergh, 2014). Furthermore, recent meta-analyses conclude that findings regarding EF in ASD are

24 inconsistent across studies. Some of the inconsistency may be due to the use of different types of outcome measures (parent-rated measures versus neuropsychological testing, the use of structured tasks versus unstructured tasks), and the inclusion of different moderators (such as age, IQ, sex). The inclusion of different diagnostic categories within the spectrum, and the presence of comorbidities that increase the risk of EF difficulties may also contribute to the inconsistencies in findings regarding EF in ASD (Geurts et al., 2014).

Some studies seem to indicate that females with ASD have more cognitive impairment compared to males, despite an apparent milder clinical presentation (Lai et al., 2015; Lemon, Gargaro, Enticott, & Rinehart, 2011). Other researchers have found that females with ASD outperform males on EF tasks (Bolte, Duketis, Poustka, & Holtmann, 2011; Lehnhardt et al., 2016). Although none of these studies have investigated the relationship between EF and social function in everyday life settings, it is important to emphasize the need to be aware of possible sex differences in ASD.

1.2.4 Measures of EF Neuropsychological testing has been the main method used to identify and quantify EF difficulties (Craig et al., 2016; Kleinhans, Akshoomoff, & Delis, 2005). However, some EF difficulties may not be observable in highly structured laboratory settings, and informant measures might therefore have higher ecological validity than neuropsychological testing (Kenworthy, Yerys, Anthony, & Wallace, 2008). For this reason, questionnaires have been developed to capture EF deficits as they occur across multiple domains of everyday life (Gioia, Isquith, Guy, & Kenworthy, 2000). Toplak and colleagues argue that performance- based measures and rating measures of EF assess different underlying constructs (Toplak, West, & Stanovich, 2013). They claim that while performance-based measures provide information of processing efficiency, rating scales provide information about goal pursuit. As a result, these two approaches to assessing EF provide different kinds of information, and do not necessarily reflect the same underlying processes or neural substrates (Toplak et al., 2013). Importantly, the most consistent and striking difficulties in people with ASD are seen on tasks that are open-ended in structure, lack explicit instructions and involve arbitrary rules (Van Eylen, Boets, Steyaert, Wagemans, & Noens, 2015; S. J. White, 2013). Individuals with ASD often display pronounced EF deficits in daily life, while performing better on highly structured neuropsychological tasks (Kenworthy et al., 2008). For this reason, I wanted to investigate EF in everyday life for children and adolescents with ASD by using a

25 questionnaire filled out by parents containing questions about everyday functioning.

The most commonly used questionnaire to investigate EF is the Behavior Rating Inventory of Executive Function (BRIEF) (Gioia et al., 2000). Based on questions regarding everyday EF, the BRIEF gives a summary score that consists of eight subscales. These subscales are divided into two indexes: the Behavioral Regulation Index (BRI) and the Metacognition Index (MI). The BRI reflects an individual’s ability to modulate both behavioral and emotional control, and to move flexibly from one activity to another. The MI is related to active problem solving, and refers to the ability to initiate, organize and monitor own actions (Gioia et al., 2000). In this way, BRI concerns appropriate behavior regulation of EF, while MI is the cognitive ability to self-manage tasks and reflect on your own performance. The common belief is that the ability to regulated behavior is a prerequisite for effective use of metacognitive processes. It is important to emphasize that metacognitive aspects of EF measured by the BRIEF are not the same at the psychological concept metacognition. The latter is defined as the control, modification and interpretations of thoughts, and is often described as “thinking about thinking” (Wells & Cartwright-Hatton, 2004).

1.2.5 The relationship between EF and intelligence EF is closely linked to fluid intelligence, which is defined as the capacity to reason and solve novel problems, independent of any knowledge from the past (Diamond, 2013). Performance on classic neuropsychological tests such as the Wisconsin Card Sorting Test, Verbal Fluency and Trail Making Test B has in some studies been shown to be entirely explained by fluid intelligence. However, for tests assessing multitasking and decision-making, EF deficits remain after controlling for fluid intelligence (Roca et al., 2014). Some researchers have found that updating/working memory is highly correlated with intelligence measures, but inhibiting and shifting are not (Friedman et al., 2006). Merchan-Naranjo et al. found significant differences in EF performance between children with ASD and a healthy control group, and these differences remained after controlling for IQ (Merchan-Naranjo et al., 2016). This might indicate that the relationship between EF and IQ is different depending on which (clinical) group is examined, and that children and adolescents with ASD have specific EF impairments that cannot solely be explained by IQ. Since IQ and EF can influence each other, it is important to take into account the IQ level of the participant when studying EF.

26 1.2.6 Comorbid ADHD and executive dysfunction Individuals diagnosed with either ASD or ADHD are characterized by EF difficulties. However, individuals in either group are described as having different executive dysfunction profiles (Matsuura et al., 2014). People with ASD often have difficulties with tasks that demand mental flexibility and verbal fluency, while children and adolescents with ADHD usually show difficulties with inhibition. Yet, inhibitory deficits can also be present in individuals with ASD (Geurts et al., 2014). Moreover, the developmental pattern of EF in children and adolescents with ASD may be atypical.

Andersen et al. found that performance on a verbal working memory task did not improve after two years in children with high-functioning ASD, while improvements were observed in children with ADHD and TD children (Andersen et al., 2015). Even though EF deficits are more closely related to the diagnostic criteria for ADHD than for ASD, it seems that EF deficits to a higher degree normalize in people with ADHD than in people with ASD (Andersen, Skogli, Hovik, & Øie, 2016). However, the focus in this thesis is children and adolescents with ASD or ASD with comorbid ADHD, and we do not compare people with ASD alone to people with ADHD alone. A recent review of EF deficits in ASD and ADHD suggests that the common co-occurrence of EF deficits seems to reflect an additive comorbidity, rather than separate conditions with distinct impairments (Craig et al., 2016). Such additive comorbidity could manifest itself as an increased amount of EF difficulties (Craig et al., 2016; Geurts et al., 2014; Pennington & Ozonoff, 1996).

People with ASD and comorbid ADHD share impairment in flexibility and planning with the ASD group, and the response inhibition deficit with the ADHD group (Craig et al., 2016). Craig et al. argue in their review that the findings are in line with the latest diagnostic criteria (DSM-5), where ADHD can co-occur together with ASD. and that comorbidity is important for treatment planning (Craig et al., 2016; Reiersen & Todd, 2008).

1.3 Social function and the relationship to EF Social function defines an individual's interactions with their environment and the ability to fulfill their role within environments such as school/work, social activities, and relationships with partners and family (Bosc, 2000). A central part of social function is social cognition, which involves how people process, store, and apply information about other

27 people and social situations. Social cognition focuses on the role that cognitive processes play in our social interactions. The diagnostic criteria for ASD emphasize difficulties with reciprocal social interaction and social communication. In addition to restricted and repetitive behaviors and interests, deficits relating to social function are at the core of an ASD diagnosis. Research indicates that social impairment is the most stable symptom of ASD across all developmental levels, and understanding other persons’ emotions, intentions, and beliefs are a necessarily prerequisite for successful social interaction (Barendse, Hendriks, Thoonen, Aldenkamp, & Kessels, 2018).

People with ASD also have a range of deficits relating to social communication. Even those with an average IQ may have difficulties using language effectively for social interaction. In addition to spoken language, communication involves all kinds of sending and receiving messages, including gestures and body language (Paul & Simmons, 2019). Social function is a very broadly defined term. It is hard to capture and assess all factors relating to successful communication and social interaction. In this thesis, I use the terms social function, social dysfunction and social difficulties in relation to deficits in social interaction and communication as defined in the diagnostic criteria for ASD.

EF difficulties in ASD have traditionally been linked to restricted and repetitive behaviors and interests (Hill, 2004). However, EF is also suggested to be important for social skills in people with ASD (Diamond, 2013). Some studies have indicated that social function and EF are independently associated with general cognitive abilities (IQ) in children and adolescents with ASD (Constantino & Gruber, 2005; Mahone et al., 2002; Vriezen & Pigott, 2002). However, few studies have investigated the relationship between EF and social difficulties in ASD (Leung, Vogan, Powell, Anagnostou, & Taylor, 2015).

There are indications that knowledge about the association between EF and social function can have an impact on identifying subgroups who may benefit from specialized cognitive interventions. Most intervention programs for children with ASD have low to moderate evidence for effectiveness, and few treatment manuals exist that specifically target cognitive dysfunctions like EF deficits for children with ASD (Lai et al., 2014). Some research suggests that treatments targeting EF may be efficient in improving both EF and social difficulties (Kenworthy et al., 2014). Children with social problems highly related to EF dysfunction may benefit more from specific training and cognitive support interventions than children with

28 social difficulties relating to other factors, such as personality variables. Some children with ASD may benefit from working memory training, while others might need cognitive flexibility training. Others will benefit from participating in social-communication training programs. Our assertion is that we need to assess the individual child’s cognitive and social profile and then tailor interventions to fit the child.

An example of an EF intervention for children with ASD is “Unstuck and On Target”. Kenworthy et al. conclude that this is an effective intervention for children with ASD. The intervention improved children’s classroom behavior, flexibility and problem-solving skills (Kenworthy et al., 2014). De Vries and colleagues investigated the effect of two EF training conditions (computerized working memory training and cognitive flexibility training) and a non-adaptive control training on children with ASD (de Vries, Prins, Schmand, & Geurts, 2015). In all three training conditions, children improved their performance on tests of working memory, cognitive flexibility and attention, but not on inhibition. An improvement in ADHD-behavior as reported by parent’s was registered over the course of the training sessions, as well. The authors conclude that it would be important to identify specific subgroups that may benefit from this type of training (de Vries et al., 2015). We will contribute to this by identifying specific subgroups of children with ASD who might benefit from specific cognitive training interventions.

Disentangling the relationship between EF and social phenotypes by using biological data in combination with clinical information will give us the opportunity to improve our understanding of ASD and could result in better diagnostic classification and interventions. Most children with ASD struggle with severe and lifelong challenges, and it is crucial to identify factors and interventions that can prevent the development of additional psychiatric problems and chronic mental health issues. Since the disorder usually manifests in early childhood, it is possible to help children get specialized and individualized help early in their development, which can help them fulfill their potential and become more active participants in society.

1.4 Unanswered questions Although it is known that EF deficits are common in ASD and important for both the understanding and treatment of the disorder, less is known about how specific EF deficits are associated to the social difficulties that characterize ASD. A large amount of research on EF

29 has been done using neuropsychological tests. This can be a limitation when trying to understand challenges in everyday life, however, because the ASD group can perform within the normal variation on highly structured tasks even though they have substantial problems with EF in everyday life. Furthermore, there is a wide variation in difficulties and functional levels in the ASD population, and there is a need to find out which subgroups are characterized by which type of difficulties. Recently, there has been an increasing interest in trying to understand whether girls with ASD have other difficulties than boys, and possible implications of this for diagnosis and treatment. Furthermore, comorbid ADHD can influence the profile of EF deficits in persons with ASD, and there is a need to understand how comorbidity impacts the link between EF and social difficulties. I am therefore also interested in investigating how sex differences and comorbid ADHD affect the relationship between EF in everyday life and social difficulties in individuals with ASD.

Although ASD is one of the most inheritable psychiatric disorders, the diagnosis is still behaviorally defined. A new and interesting approach to investigating ASD is to look at PGS and how such scores relate to EF deficits and social function. I wanted to investigate the relationship of several PGSs to EF deficits in ASD. By exploring the association between EF deficits and social difficulties on different levels of explanation, from genetic tools like PGS to clinical characteristics and questionnaires assessing function in everyday life, I aimed to generate new knowledge that can contribute to understanding the needs of children and adolescents with ASD. This information could be relevant for the stratification of treatment.

2. Aims The main aim of this thesis is to investigate the relationship between EF and social difficulties in children and adolescents with ASD by integrating different levels of knowledge – genetics, clinical symptom characteristics, and everyday function reports. Furthermore, I intend to examine how sex differences and comorbid ADHD influence the relationship between EF and social difficulties in this group of patients. The thesis includes the following sub-aims:

1. To investigate the association between social function as measured by the Social Responsiveness Scale (SRS) and everyday EF measured by the BRIEF in a clinical sample of children and adolescents with ASD. I also investigate potential sex differences and possible age differences (Paper I).

30 2. To examine the relationship between EF in everyday life as rated by parents (using BRIEF) and assessment of autistic symptomology (using ADI-R) , and to investigate possible sex differences in this relationship (Paper II). 3. To explore if the polygenic scores (PGSs) for ASD, ADHD and general intelligence (INT) are associated with EF in everyday life using BRIEF in a clinical sample of children and adolescents admitted for clinical assessment of ASD. In addition, I investigate whether the PGSs for ASD, ADHD and general intelligence (INT) are associated with social function in everyday life (SRS) (Paper III).

3. Methods 3.1 Design The studies applied a naturalistic, cross-sectional design. The patient sample consisted of children and adolescents referred to different child and adolescence mental health centers for ASD evaluation in Norway. Overlapping clinical subgroups, with a special focus on comorbid ADHD, were used to test the specificity of the findings. In addition, 29 typically developed (TD) Norwegian controls were included in Paper III. The project had access to large international PGS samples through the international Psychiatric Genomics Consortia (Demontis et al., 2018; Grove et al., 2019; Savage et al., 2018) .

3.2 Participants The study is part of the national BUPGEN network, recruiting patients from the Norwegian Health Service specializing in the assessment of ASD and neurodevelopmental disorders. The main BUPGEN inclusion criterion was that either the 1st line or 2nd line of health services had referred the child to a mental health center for a diagnostic assessment of ASD. Participants did not need to receive a diagnosis in the ASD spectrum, and they could have other comorbid developmental disorders. Participants included in the studies that form the bases of this thesis had an IQ within the normal range based on standardized intelligence measures (IQ ≥ 70), spoke Norwegian fluently, and had a complete data set from the required assessments (BRIEF, SRS, ADI-R and PGS). Participants could have a comorbid ADHD diagnosis. Exclusion criteria were significant sensory loss (vision and/or hearing) that could interfere with testing. The recruitment was ongoing, and this thesis includes data from the BUPGEN database in 2018 (Paper I), 2018/19 (Paper II), and 2019 (Paper III). Paper I included 86 participants, Paper II included 116 participants (66 participants in the FUZZY

31 matched sample), and Paper III consisted of 176 participants. For a description of sex distribution, age, IQ and ASD diagnoses of the participants included in each of the papers, see Table 1-4. I recruited about 80 participants from Bærum BUP, where I was working throughout my PhD period.

Table 1. Sample 1. Paper I. Recruitment period January 2013 – December 2016

N 86

Girls/ boys 23/ 63

Age, mean (SD) 13.0 (2.7) Range 6-18

ASD subgroups N Childhood autism (F84.0) 13 Atypical autism (F84.1) 1 Asperger syndrome (F84.5) 41 PDD-NOS (F84.9) 31

Comorbid ADHD 28

Full-scale IQ (SD) Range 93.4 (14.5) Range 71-133

Table 2. Sample 2. Paper II. Recruitment period January 2013 – May 2018

N 116

Girls/ boys 25/ 91

Age, mean (SD) Girls 12.0 (3.1)/ Boys 10.4 (3.2) Range 5-19

ASD subgroups N (girls/boys) Childhood autism (F84.0) 15 (2/ 13) Atypical autism (F84.1) 9 (2/ 7) Asperger syndrome (F84.5) 57 (14/ 43) PDD-NOS (F84.9) 35 (7/ 28)

Comorbid ADHD 40 (5/ 35)

Full-scale IQ (SD) Range 93.5 (9.3)/ 95.6 (13.1)

32 Table 3. Sample 2. Paper II. FUZZY matched sample. Recruitment period January 2013 – May 2018

N 66

Girls/ boys 22/ 44

Age, mean (SD) Girls 11.9 (3.1)/ Boys 11.2 (3.3) Range 5-19

ASD subgroups N (girls/boys) Childhood autism (F84.0) (girsl/boys) 9 (2/ 7) Atypical autism (F84.1) 5 (2/ 3) Asperger syndrome (F84.5) 27 (11/ 16) PDD-NOS (F84.9) 25 (7/ 18)

Comorbid ADHD 17 (4/ 13)

Full-scale IQ (SD) Range Girls 93.5 (9.3)/ Boys 92.1 (11.3)

Table 4. Sample 3. Paper III. Recruitment period January 2013 – December 2018

N 176

Girls/ boys 42/ 134

Age, mean (SD) 11.7 (3.7) Range 5-22

ASD subgroups (n=120) Childhood autism (F84.0) 18 Atypical autism (F84.1) 5 Asperger syndrome (F84.5) 56 PDD-NOS (F84.9) 41

Comorbid ADHD 42 of 120 ADHD without ASD 26 of 56

Full-scale IQ (SD) Range 93.7 (13.5) Range 70-133

3.3 Procedures The participating children and adolescents underwent a standard assessment according to established clinical practice guidelines. The clinical evaluations and measurements included in this study were carried out on an ongoing basis. A team of experienced clinicians (clinical psychologists and/or child and adolescence psychiatrist) assessed the children and adolescents. Diagnostic conclusions were best-estimate clinical diagnoses derived from tests, interview results and observations. All diagnoses were based on ICD-10 criteria (World Health Organization, 1992). All of the participants followed the Standard Operating Procedure of the BUPGEN project. In relation to the questionnaire used in the study, we used the parent version where most often the questionnaires were filled out jointly by both parents. If the parents did not fill out the questionnaire together, we prioritized the mother’s response over the father’s response.

33

3.4 Measures 3.4.1 Clinical assessment Participants underwent a clinical evaluation for ASD using at least one of the gold standard tools the Autism Diagnostic Observation Schedule (ADOS), or the Autism Diagnostic Interview-Revised (ADI-R), or the Social Communication Questionnaire (SCQ) evaluated by an experienced clinician (Lord et al., 2000; Lord, Rutter, & Le Couteur, 1994; Michael Rutter, Bailey, & Lord, 2003; M. Rutter, Lord, & LeCouteur, 2016). In addition, a comprehensive diagnostic assessment was conducted based on a full medical and developmental history, questionnaires, clinical interviews, naturalistic observations and formal testing of cognitive function.

3.4.1.1. Autism Diagnostic Interview-Revised (ADI-R) The ADI-R diagnostic algorithm was used to assess autistic symptoms. The ADI-R is a clinical diagnostic tool based on a comprehensive interview with parents or primary caregivers of the child/ adolescent (Lord et al., 1994). The interview consists of 93 questions, and a predetermined number of these scores are entered into a diagnostic algorithm. The interview and scoring follow standardized procedures, and the interviewer records and codes the informant’s responses. The algorithm is divided into three functional domains based on the diagnostic criteria (qualitative deviations in): A = Reciprocal Social Interaction, B = Communication, C = Restricted, Repetitive, and Stereotyped Behavior. Higher scores indicate that an individual has a greater number of items representing core ASD deficits and/or more severe symptoms (Gotham, Pickles, & Lord, 2009). All the participants were verbal children, and therefore the algorithm for verbal children was used. We used the Norwegian translation of the ADI-R (M. Rutter et al., 2016).

3.4.1.2 Autism Diagnostic Observation Schedule (ADOS) The Autism Diagnostic Observation Schedule is a clinician-administrated diagnostic tool to assess ASD behavior and symptoms. The ADOS consists of different modules based of the age and language abilities of the child/ adolescent/ adult, and involves the administration of a clinically administered observation assessment and standardized activities. Items are rated on a scale of 0-3, with higher scores indicating greater impairment (Lord et al., 2000). ADOS

34 scores and autism core symptoms are found to be remarkably stable over time across childhood (Bieleninik et al., 2017)

3.4.1.3 Cognitive testing/ intelligence The intelligence quotient (IQ) was measured using the Wechsler’s Intelligence Scale for Children - Fourth Edition (WISC-IV) (Wechsler, 2003) or other Wechsler test for the appropriate age group (Wechsler, 2002, 2008). A full-scale IQ score of 100 represents the average score, and the standard deviation is +/- 15 points. In all the studies, participants had a full-scale IQ score of 70 or above.

3.4.2 Questionnaires 3.4.2.1 Behavior Rating Inventory of Executive Function (BRIEF) The BRIEF is one of the most commonly used clinical tools to measure EF in everyday life (Gioia, Isquith, Guy, & Kenworthy, 2015). The questionnaire is designed to assess EF in everyday life as rated by parents, teachers and/or self-report. BRIEF consists of 86 questions which are rated on a 3-point Likert scale (Gioia et al., 2000). In all the papers, we used the t- scores from the parental reports for the age group 5-18 years. The BRIEF consists of three indexes and eight subscales. The Behavioral Regulation Index incorporates the three subscales: inhibit, shift and emotional control. The Metacognition Index consists of the five subscales: initiate, working memory, initiate, plan/organize, organization of materials and monitor. Together, these two indexes, the Behavioral Regulation Index (BRI) and Metacognition Index (MI), form the Global Executive Composite Index (GEC). T-scores ≥65 are considered to represent clinically significant areas. The internal consistency is reported to be high (Cronbach’s alpha = 0.80 – 0.98) (Gioia et al., 2000). For an overview and description of BRIEF indexes and subscales, see Table 5.

3.4.2.2 Social Responsiveness Scale (SRS) To identify social difficulties, we used the parent version of the questionnaire SRS for the age group 4-18 years. SRS has 65 questions rated on a 4-point Likert scale. This continuous scale allows trait quantification that many diagnostic tools fail to capture (Achenbach, 2011; Nelson et al., 2016). In addition to a total score, SRS consists of five treatment subscales: Social Awareness, Social Cognition, Social Communication, Social Motivation and Autistic Mannerisms. The SRS has been translated into Norwegian (Ørbeck, 2009). A higher score

35 indicated more ASD symptoms (t-score under 59 mild symptoms, t-score 60-75 moderate symptoms, t-score 76 or higher severe symptoms). Reliability is high both in standard and clinical samples (r = .93 to .97) (Constantino & Gruber, 2005). The SRS’s reliability and validity have been proven to be satisfactory in both population-based and clinical samples in Europe and in the USA (Bolte, Poustka, & Constantino, 2008; Wigham, McConachie, Tandos, Le Couteur, & Gateshead Millennium Study core, 2012). The scale correlates well with the scores in the Autism Diagnostic Interview-Revised (ADI-R) (Constantino et al., 2003). Most findings support a one-factor model of SRS (Bolte et al., 2008; Constantino et al., 2004), while some have found evidence for several dimensions (Nelson et al., 2016).

Table 5. Behavioral description of the clinical scales in Behavior Rating Inventory of Executive Function (BRIEF)

BRIEF Clinical Scales Behavioral description Behavioral Regulation Index (BRI) Inhibit Control impulses, appropriately stop own behavior at the proper time Shift Move freely from one situation, activity, or aspect of a problem to another as the situation demands, transition, solve problems flexibly Emotional Control Modulate emotional responses appropriately Metacognition Index (MI) Initiate Begin task or activity, independently generate ideas Working Memory Hold information in mind for the purpose of completing a task, stay with, or stick to, an activity Plan/ Organize Anticipate future events, set goals, develop appropriate steps ahead of time to carry out associated task or action, to carry out tasks in a systematic manner, understand and communicate main ideas or key concepts Organization of Materials Keep workplace, play areas, and materials in an orderly manner Monitor Check work, assess performance during or after finishing a task to ensure attainment of goal, keep track of the effect of own behavior on others Global Executive Composite (GEC) Behavioral Regulation Index + Metacognition Index (altogether 8 subscales)

Adapted from Gioia (Gioia et al., 2000).

3.4.2.3 Social Communication Questionnaire (SCQ) The SCQ was developed as a screener to evaluate communication skills and social functioning in children who may have ASD. The SCQ is based on ADI-R, and the questions were deliberately chosen to match the ADI-R items that were found to have discriminative diagnostic validity (Michael Rutter et al., 2003). We used the Lifetime Form, which consists of 40 yes-or-no questions; a parent or other primary caregiver completed the questionnaire. The recommended cut-off score on the Lifetime Form is 15 or greater, and this might be an indication of possible ASD and the need for a more comprehensive evaluation. The internal consistency is reported to be acceptable to good and increasing with age (Cronbach’s alpha = 0.84 – 0.93) (Michael Rutter et al., 2003).

36 3.4.3 Polygenic score (PGS) Genotyping Participants’ DNA was extracted using standard procedures from blood or saliva samples collected in the clinic. Both clinical and control participants’ samples were collected in the same time period, chip-genotyped on the same array, and imputed and processed together. All DNA samples were chip-genotyped on Human Omni Express-24 v.1.1 (Illumina Inc., San Diego, CA, USA) at deCODE Genetics (Reykjavik, Iceland). The chip-genotypes were quality-controlled using PLINK 1.9 (Purcell et al., 2007). In short, genetic variants were excluded if they had low coverage (< 95%) or minor allele frequency (MAF) (< 0.01), deviated from Hardy-Weinberg equilibrium (p-value < 0.0001), or had significantly different prevalence in different genotyping batches (FDR < 0.5). Entire individual genotypes were excluded if they had low coverage (less than 95%) or a high likelihood of contamination (heterozygosity above mean + 5 standard deviations). No relatedness above pi-hat 0.1 was detected in our sample. MaCH software was deployed to impute the participants’ genotypes using the reference haplotypes from the European population in the 1000 Genome Project - phase III (Genomic build 37) (Das et al., 2016; Li, Willer, Ding, Scheet, & Abecasis, 2010). Any imputed variants with MAF lower than 0.05 or an information score lower than 0.8 were excluded from the subsequent PGS calculations.

Polygenic score (PGS) The variants’ effects were estimated from an inverse-variance weighted meta-analysis of the GWAS summary statistics for all original sub-studies, with the exception of any sub-studies our own participants were drawn from (BUPGEN). All summary statistics were quality- controlled by excluding variants that met any of the following conditions: MAF < 0.05; imputation quality INFO < 0.8; or present in less than half of the sub-studies. The remaining variants were clumped into independent regions on the basis of the linkage disequilibrium structure of the 1000 Genomes Phase III European population using PLINK v1.9 (with options: --clump-p1 1.0 --clump-p2 1.0 --clump-r2 0.2 --clump-kb 500). The allelic dosage coefficient (or logarithm of the odds ratio) of each of the variants with minimum p-values from all independent regions were taken as weights in constructing the polygenic scores. We obtained GWAS summary statistics for ASD from Grove et al. (2019), ADHD from Demontis et al. (2018) and INT from Savage et al. (2018) (Demontis et al., 2018; Grove et al., 2019; Savage et al., 2018). The GWAS summary statistics were used to generate PGS and to test the association with specific phenotypes in the current sample following the procedure referred to

37 as “genetic risk profiling” (Martin, Daly, Robinson, Hyman, & Neale, 2018). The participants in our subsample, as well as those in the Grove et al. (2019), Demontis et al. (2018), and Savage et al. (2018) studies, are all of European ancestry and therefore expected to be genetically relatively homogeneous. The discovery samples our PGSs are based on consisted of 18,381 ASD cases and 27,969 controls for the ASD PGS (Grove et al., 2019), 20,183 ADHD cases and 35,191 controls for the ADHD PGS (Demontis et al., 2018) and 269,867 cases for the INT PGS (Savage et al., 2018).

3.4.4 Copy-number variation (CNV) As mentioned earlier, CNVs have in molecular studies been identified as risk factors for ASD (Geschwind, 2011; Sullivan & Geschwind, 2019). Information about the CNVs in our sample was obtained from the chip genotypes, following standard CNV calling methodology (Sonderby et al., 2018).

3.5 Statistical analyses The statistical analyses were conducted using the statistical package IBM SPSS Statistics for windows, version 23.0 (Paper I) and 25.0 (Paper III). The analyses in Paper II were conducted using the R statistical environment (version 3.5.0.), the “jmv” (version 0.7.3.1.) and the “cocor” packages. Statistical significance was set at p <0.05. To control for Type-I errors, we adjusted our alpha level in Paper I and Paper II. P-values are therefore interpreted differently depending on the test and the number of comparisons done. P-values less than 0.05 and higher than the adjusted p-level are referred to as “tendencies” in the text. Conventional values were used for interpreting effect sizes (0.2 small, 0.5 medium and 0.8 large (Cohen, 1988)). In the regression analyses, assumptions of linearity, multicollinearity, independence of errors, homoscedasticity, unusual points and normality of residuals were met.

In Paper I, a Pearson’s independent t-test was used to compare means (girls/boys, age groups); Chi-square for crosstabs was applied to investigate differences in distributions (ASD and ADHD diagnoses). We used the online calculator http://vassarstats.net/rdiff.html (two- tailed) to evaluate if the subgroups’ correlation coefficients (girls/boys, age groups) significantly differed. Separate multiple regression analyses were conducted to explain the variance in SRS scores (total and subscales) from age, sex, total IQ and BRIEF (BRI and MI subscales). A Bonferroni correction was used to correct the significance level for multiple

38 comparisons: p < 0.01 was used in all the regression models and p <0.003 was used in the correlation analyses. In all other analyses (t-tests, Chi-square and differences in correlations coefficients), we applied p < 0.01.

In Paper II, we used the Welch test to assess sex differences in ADI-R and BRIEF scores. The Welch test was used because it is robust to unequal variances. We adjusted the p-level for 6 tests and sat the critical significance level at p < 0.008. Glass’ delta, which is unaffected by unequal variances, was used as a measure of effect size. Then, we used Fisher’s z test to compare if the correlation coefficients between the BRIEF GEC and ADI-R sub-scores were significantly different between the girls and the boys. To examine the frequency distribution of comorbid ADHD between the sexes, we used chi-squared statistics. We fitted nested linear regression models to assess the role of covariates (sex, IQ, age and ADHD diagnosis) and the interaction of sex and EF on the relationship between ADI-R scores and BRIEF GEC scores. We did this separately for the three ADI-R sub-scores: reciprocal social interaction, communication, and restricted, repetitive and stereotyped behaviors. In the first model, we incorporated sex, IQ, ADHD diagnosis and age. In the second model, we added BRIEF GEC in addition to the previously mentioned variables. In the third model, we added the interaction term BRIEF GEC X sex. Then, we compared the fit of these models by calculating Akaike information criterion (AIC) values and F-ratios for the model change. Lower AIC values are indicative of a better model fit. We adjusted for three tests, and the significance level was set at p <0.017. To generalize the regression results beyond the given samples, robust regression was performed in the event of non-normally distributed standardized residuals bootstrapping with 2000 samples. We obtained bootstrapped 95% confidence intervals for the model intercept and slopes, and compared these with the confidence intervals in the original model. To examine the impact of the more closely matched boys and girls on age and IQ, the same model fit and comparison procedures were performed on a subset of the sample, which was generated using the FUZZY extension command in SPSS. We matched every girl to two boys (1:2) on age within 2 years and total IQ within 10 points.

In Paper III, we divided all the clinical participants into the three groups: low, moderate and high, based on each of their PGSs for ASD, ADHD and INT (altogether nine groups). The groups were identified based on the normal distribution curve of the standardized PGS scores. The moderate group consisted of all participants with PGS within one standard deviation from the mean. The low and high groups consisted of participants with PGS in the lower and upper

39 tails of the distribution, respectively. This subdivision roughly corresponds to the participants with the 15% lowest scores constituting the Low group, those with the 15% highest scores making up the high group, and everyone in between in the moderate group. Pearson’s independent t-test was used to investigate the mean difference between all of the clinical variables between the low and high PGS groups; and, to examine the differences in PGS between controls and the clinical group in order to validate the ASD PGS in our sample. Chi- square for crosstabs was used to investigate differences in the distribution of sex and ASD, and ADHD diagnoses in the low and the high ASD PGS groups. Separate linear regression models were applied to explain variance in BRIEF scores (BRI, MI and GEC) across ASD, ADHD and INT PGS groups. Age, sex, total SRS score, full-scale IQ and principal components (PCs) were entered as covariates. We also performed linear regression with total SRS score as the dependent variable, and ASD, ADHD INT PGS groups, age, full-scale IQ, BRIEF scores (BRI and MI), and PCs as independent variables. As the sample size did not permit inclusion of all PC’s, the genetic PCs that were entered into the linear regression model were the ones with the highest correlations to the PGSs (ASD, ADHD and INT) and the dependent variables (BRIEF GEC and SRS), respectively. We also fitted the regression models with continuous PGS instead of group PGS. Due to a moderate sample size and the explorative nature of this paper, we did not adjust alpha levels for multiple testing because of the risk of type 2 errors. Therefore, all tests were considered significant with a two-sided p- value lower than 0.05.

3.6 Ethical considerations The study was approved by the Regional Ethical Committee and the Norwegian Data Inspectorate (REK #2012/1967), and was conducted in accordance with the Helsinki Declaration of the World Medical Association Assembly. The study is based on assessments and tests that are used in regular clinical practice. The participants, their families and the professional network around them were offered information about the test results, the diagnostic process and recommended interventions. Thus, information learned in the study was used to in the clinic to enhance assessment and treatment of the participants. In addition, we collected DNA from either a blood or saliva sample. The participants did not get the results from the research DNA test. If it was necessary to perform some genetic testing as part of the clinical assessment, this was done in addition to the research test, and the patient/family got the results from these tests. The patient and/or family could withdraw their participation in the project at any time, and this did not influence their right to clinical help or follow-up

40 treatment. Written, informed consent was obtained from parents or a legal guardian for participants under 18 years. We informed all participants about inclusion in the study, and adolescents between 12-17 years of age were encouraged to co-sign the informed consent form. Participants over 18 years gave written informed consent themselves.

Through the BUPGEN study and a collaboration with the Norwegian Autism Society, we described and explored the attitudes of a total of 1455 parents of children with ASD with regard to clinical genetic testing and genetic research. The majority of the parents were positive to genetic testing. Data confidentiality and information about the purpose and progress of the research were particularly important for them regarding genetic research, and the possibility of finding etiological explanations was their main motivation in relation to clinical genetic testing (Johannessen et al., 2016; Johannessen et al., 2017).

As of today, no specific EF treatment is offered to children and adolescents with executive dysfunction in our clinic. Instead, the child, their parents and their teachers are given information about executive dysfunction and how this can affect the child in everyday life at school, at home and when participating in activities. The information is used to discuss how to adapt the school and home environments to the child’s challenges, and how the child can practice improving his or her EF skills in everyday activities. Furthermore, PGS is at the present time not used in the clinic. The introduction of PGS into clinical practice will be associated with some ethical challenges relating to communicating uncertain findings to patients. Therefore, care must be taken to ensure that the information, interpretation and understanding of possible risk factors are communicated sensitively. Ultimately, evidence- based clinical translation of PGS can lead to more effective diagnosis, treatment and possibly even prevention of particular psychiatric disorders (Palk et al., 2019).

4. Summary of papers Paper I: Metacognitive aspects of executive function are highly associated with social functioning on parent-rated measures in children with autism spectrum disorder. Background: EF deficits are common in children with ASD, but not part of the diagnostic criteria. While both EF deficits and social difficulties in ASD have been studied extensively separately, there has been limited research on the relationship between EF and social function. Knowledge about the relationship can have implications for the understanding of cognitive

41 components related to the social difficulties that define ASD and may be relevant for developing more targeted EF interventions for people with ASD. Aims: The aim of the present study was to investigate the relationship between social functioning measured with the SRS and everyday EF measured with the BRIEF in a clinical sample of children and adolescents with ASD. We also investigated potential sex differences and possible age differences by splitting the sample into two age groups (6-12 years and 13- 18 years). Method: The study included 86 children and adolescents (23 girls) with ASD (age 6-18). Of these, 28 participants had a comorbid diagnosis of ADHD. All participants had an IQ within the normal range based on a standardized Wechsler’s test (Total IQ ≥ 70) and spoke Norwegian fluently. The parent version of the SRS was used to identify social communication difficulties. The BRIEF parent version was used to assess EF deficits in everyday life situations. Results: The main finding of the current study was that metacognitive aspects of EF (The Metacognition Index) was the most important factor in explaining social function in children with ASD when controlling for other confounders like sex, age, and full-scale IQ (p = 0.001). We did not find any significant differences between the girls and the boys in the relationship between EF deficits and social difficulties in our sample, yet there was a tendency for a stronger relationship for the girls compared to the boys. Nor did we find any significant differences between the two age groups children (6–12 years) and adolescents (13–18 year). The results revealed that the oldest age group only had a significant relationship between social function and metacognitive aspects of EF (r = 0.47, p = <0.001), but not between social function and behavior regulation aspects of EF (r = 0.34, p = 0.015). This may imply that metacognitive EF is of particular importance for social function in older age, and that behavior regulation aspects of EF do not have the same impact on social function for the oldest age group. Conclusion: We report a relationship between parental reports of EF deficits and social difficulties in an everyday setting in children with ASD. We found that metacognitive aspects of EF had a significant association to most aspects of social function. Further studies are needed to clarify if children with ASD will improve their social function through intervention programs designed to enhance EF in general and metacognitive function of EF in particular.

42 Paper II: Sex differences in the relationship between social difficulties and executive dysfunction in children and adolescents with autism spectrum disorder. Background: The prevalence of ASD in boys is nearly four times higher than in girls, and the causes of this sex difference are not known. EF difficulties may play a part in the development of autistic symptomatology. In this study, we investigated sex differences in the relationship between EF in everyday life (parent-rated) and social dysfunction symptoms in a representative sample of ASD children. Aim: The aims of the study were to examine the relationship between EF in everyday life as rated by parents of children and adolescents with ASD and autistic symptomology, and to investigate possible sex differences in this relationship. Method: The study included 116 children (25 girls) aged 5-19 years with a verified diagnosis of ASD. The participants’ full-scale IQ was above 70, and they were assessed with the BRIEF and the ADI-R. Results: There was a statistically significant correlation between two ADI-R domains (reciprocal social interaction and communication) and the BRIEF total score for girls (r = 0.70-0.78), while no significant correlations were found for boys (r = 0.18-0.23). The identified correlations were significantly different between sexes for each ADI-R domain (p <0.001-0.010). Regression models showed that comorbid ADHD did not contribute significantly to the relationship between BRIEF and ADI-R scores. Conclusion: The results demonstrate a relationship between difficulties with social reciprocity and communication and parent-rated executive dysfunction in girls with ASD, but this association was not significant in boys with ASD. We did not find sex differences in the relationship between executive dysfunction and restricted and repetitive behaviors. This suggests that there are sex differences in the relationship between EF deficits and social difficulties in ASD. Altogether, our results provide a greater understanding of how sex differences impact the clinical manifestation of ASD and may have implications for the development of sex-specific interventions for children and adolescents with ASD.

Paper III: Autism spectrum disorder polygenic scores are associated with every day executive function in children admitted for clinical assessment. Background: To better capture the polygenic nature of complex disorders, PGS can be calculated from large genome-wide association studies (GWAS) and used to estimate the sum of genetic risk in a given sample. ASD in a highly heritable disorder and many people with ASD have executive dysfunction. We wanted to investigate the relationship between different PGSs and their relationship to EF.

43 Aim: The aim of the study was to explore in a clinical population of children and adolescents assessed for ASD the association between PGSs for ASD, ADHD and general intelligence (INT) and EF deficits in everyday life. Method: The study included 176 children, adolescents and young adults aged 5-22 years with full-scale IQ above 70 and who were admitted for clinical assessment for ASD symptoms. We investigated how the three PGSs for ASD, ADHD and general intelligence (INT) relate to EF in everyday life (as measured by the BRIEF). Furthermore, we investigated the relationship between these three PGSs and social function measured with the SRS. Results: In a regression model, the ASD PGS was significantly related to behavior regulation aspects of EF (p = 0.018). We did not find significant associations between PGSs for either ADHD or INT and behavior regulation aspects of EF (p = 0.205-0.732). A high PGS for INT was significantly related to social difficulties in everyday life (as measured by the SRS). Conclusion: The common ASD risk gene variants had a stronger association to behavior regulation aspects of EF than the PGSs for ADHD and intelligence in a clinical sample of children and adolescents with ASD symptoms. Furthermore, a high PGS for general intelligence was related to social problems. If children with a genetic risk for ASD can be identified, then this could be used in the future to initiate interventions aimed at EF deficits, or to stratify the more general ASD treatment approach according to PGS.

5. Discussion

5.1 Main findings 5.1.1. Summary of main findings Summary of main findings related to the associations between parent-rated executive dysfunction and social difficulties in children and adolescents with ASD:

• The metacognitive aspects of EF were significantly related to everyday life social functioning in children and adolescents with ASD. This result remained after excluding the participants with comorbid ADHD from the analyses. There were no significant differences in the relationship between EF and social function between boys and girls, or the youngest (6-12 years) versus the oldest group (13-18 years). However, there were tendencies for a stronger relationship between EF deficits and social difficulties for girls and the youngest age group.

44 • There was a significant association between EF difficulties in everyday life and autistic symptoms related to communication and social interaction for girls with ASD, but not for boys with ASD. We did not find sex differences in the relationship between executive dysfunction and restricted and repetitive behaviors.

• There was a significant association between a high ASD PGS and EF deficits in everyday life relating to behavior regulation. The PGSs for ADHD and general intelligence did not contribute significantly to EF. Sex had a significant contribution in the regression analysis, and the association between the ASD PGS and the behavior regulation EF seemed to be stronger for boys than for girls. We used genetics and PGS as a tool to investigate the relationship between social function and EF, and the association between the two is reinforced by the fact that we find a connection between ASD biology and EF in everyday life. Furthermore, we found social function, as measured by the SRS scale, to be related to the PGS for general intelligence.

5.1.2 The relationship between metacognitive aspects of EF and social difficulties In our first paper, we found that metacognitive aspects of EF were particularly important for social function in the children and adolescents with ASD in our sample. Metacognitive aspects of EF are related to the child’s ability to engage in active problem solving, and to use working memory to initiate, organize and monitor their own actions (Gioia et al., 2000). This process puts demands on the child’s capacity to cognitively self-manage tasks and to reflect and monitor his/her own actions. In everyday life, these kinds of difficulties may present as trouble getting started with tasks/homework, being overwhelmed by large assignments, needing help from an adult to stay on task or being unaware of how his/her behaviors bother others (Gioia et al., 2000).

The finding that metacognitive aspects of EF are important for social function is in line with previous studies (Kenworthy, Black, Harrison, della Rosa, & Wallace, 2009; Leung et al., 2015). However, theses previous studies also found that the behavior regulation part of EF had a significant contribution to at least some parts of social function. The deviation in findings might be the result of different measures being used, where Kenworthy et al. used a composite score of ADOS and ADI-R instead of the SRS. We also used a stricter significance level/ alpha level than Kenworthy et al. and Leung et al. Furthermore, we analyzed our findings on a more detailed level, by looking at the relationship between different parts of EF

45 and specific aspects of social function. In particular, difficulties relating to social cognition, social communication and autistic mannerisms appeared to have a stronger relationship to EF in the correlation analyses than the ability to notice social clues and social motivation. In the regression analysis, however, the metacognitive aspects of EF were a significant predictor of the subscales social communication, social motivation, autistic mannerisms and the SRS total score. The behavior regulation aspects of EF did not provide a significant contribution to any of the subscales from SRS. The finding regarding the subscales from SRS must be interpreted with caution, since most findings support a one-factor model of SRS (Bolte et al., 2008; Constantino et al., 2004). In general, it is important to recognize that although participants had most difficulty (highest scores) with behavior regulation aspects of EF, metacognitive aspects of EF had the strongest association to social difficulties. Deficits relating to metacognitive aspects of EF, such as problems with working memory, planning and monitoring, might be easier to overlook by parents and teachers, than difficulties with behavior regulation, which is more directly observable. It is therefore important to emphasize metacognitive difficulties and their potential relationship to social dysfunction.

Some researchers conceptually distinguish between hot and cold EF. Hot EF involves motivational and emotional aspects of decision-making e.g. delayed discounting and affective decision-making. On the other hand, cold EF consists of inhibition, working memory and planning in more neutral and nonaffective situations (Kouklari, Tsermentseli, & Monks, 2019; Zimmerman et al., 2016). When it comes to the questionnaire BRIEF, the Metacognition Index (MI) is thought to be linked to cold EF, and the Behavioral Regulation Index (BRI) is linked to hot EF (Puhr et al., 2019). Furthermore, EF might influence the emergence and development of ToM, and the expression of ToM (Carlson, Claxton, & Moses, 2015; Pellicano, 2010). Even though some have found hot EF to be related to ToM over and above the effect of cold EF, most researchers have found cold EF to be closely connected to ToM (Kouklari et al., 2019). Thus, deficits in metacognitive aspects of EF, which is mostly cold EF, may be closely linked to core difficulties in ASD such as ToM and perspective-taking. In general, successful social interaction and communication puts heavy demands on a person’s ability to reflect and monitor their own actions, to take another person’s perspective and to understand when their own behavior can bother others. These are all areas in which people with ASD have challenges and might represent different aspects of the same core difficulty with social reciprocity. For people with ASD, who from the outset have problems with social communication and reciprocity, it may be that metacognitive EF becomes particularly

46 important for mastering and understanding what is required in social settings. Leung et al. argue that the widespread use of EF skills is probably not necessary for successful social interaction for TD children with normally developed EF, because these children have a more efficient use of EF and better social competence than children with ASD (Leung et al., 2015). Interestingly, the metacognitive aspects of EF, cold EF and ToM may be interconnected, mutually affect each other, and play a central role in the social dysfunction of people with ASD.

While only the metacognitive aspects of EF were significantly correlated to social function in the oldest age group in our study (13-18 years), both behavior regulation and metacognitive aspects of EF were significantly correlated to social function in the younger group (6-12 years). This finding might be related to a greater degree of common EF factors in children/adolescents and the presence of both unity and diversity among older subjects (Karr et al., 2018). In this way, it may be that more common EF factors are related to social function in the youngest children in our study, while, as the children get older, more specific EF components become associated with social function.

Chouinard et al. have recently done a follow-up study where they investigated the relationship between metacognitive aspects of EF and social function in male and females both with and without ASD (Chouinard, Gallagher, & Kelly, 2019). Their study design was somewhat different from ours in that they only included participants with IQ over 85 and they used four SRS subscales to create what they called an SRS social score. They found that for both sexes metacognitive EF had explanatory value for social function, and for males this relationship was stronger for the ASD group than the non-ASD group. They concluded that females with and without ASD have a similar relationship between everyday EF and social function, while for males the relationship was different for those with and without ASD. Furthermore, in males they found the behavior regulation aspects (BRI) of EF to be a predictor of social function, while this was not the case for females. This is in contrast to our study where we found significant correlations between both BRI and MI and social function in males and females. In our regression analyses, however, only MI was significantly associated with social function and sex showed strong tendencies towards being significant but did not remain after correction for multiple comparisons. In this way, both studies confirm the importance of metacognitive EF for social function and that this relationship can be affected by the sex of the individual. In our study, we only included people with ASD. Since Chouinard et al. also

47 included participants without ASD, they could elaborate on sex differences in the relationship between EF and social function and how girls with and without ASD showed a more similar pattern than boys with and without a diagnosis. I will discuss sex differences in the relationship between EF and social difficulties in greater detail in section 5.1.3.

Our results did not show an association between high PGS for ASD and the metacognition aspects of EF (Paper III). So even though our earlier study (Paper I) showed that metacognitive aspects of EF are strongly associated with social function in ASD, we did not find the same relationship between metacognitive aspects of EF and common genetic variants for ASD. However, we did find an association between behavior regulation aspects of EF and high PGS for ASD. This will be discussed in more detail in section 5.1.4.

Several researchers have found the metacognitive aspects of EF measured with BRIEF to be of particular importance to adaptive function in children with ASD. Adaptive function includes socialization, communication and activities of daily living (ADL), and Vineland-II is the most commonly used instrument to quantify adaptive skills (Chatham et al., 2018). Gilotty et al. found that the subscales Initiate and Working Memory from the BRIEF were strongly (negatively) associated with the domain scores Communication and Socialization from Vineland-II (Gilotty et al., 2002). More recent research has confirmed that metacognitive processes represent the most significant contribution to ADL, as well as to communication and social skills (Gardiner & Iarocci, 2018). Since SRS measures social function and autistic symptoms in an everyday setting, there might be a strong resemblance to adaptive function measures like the Vineland-II. Understanding that metacognitive functions of EF are of particular importance for social function in children and adolescents with ASD may be particularly important when targeting interventions aimed at improving everyday function for this group. Gardiner & Iaroccis argue that there is great potential for measurements of everyday EF to target interventions for children both with and without ASD (Gardiner & Iarocci, 2018). Others have also found metacognitive EF training to be a promising intervention for young children with ADHD (Tamm & Nakonezny, 2015). Metacognitive EF interventions may thus be a potential intervention strategy for several neurodevelopmental diseases and/or comorbid conditions.

In summary, metacognitive aspects of EF seem to be especially important for social function in the everyday life of children and adolescents with ASD. Others have found that this

48 association seems to be particularly strong in children and adolescents with ASD compared to TD (Chouinard et al., 2019). To some degree, there might be an association between these metacognitive aspects of EF and cold EF, and further to ToM. However, our study did not investigate the relationship between these three concepts. We did not find an association between PGS for ASD and metacognitive aspects of EF, so the polygenetic aspects of ASD are not likely to be the cause of the relationship between metacognitive aspects and social difficulties. However, the metacognitive parts of EF have been linked to adaptive function in ASD in other studies (Gardiner & Iarocci, 2018). Overall, metacognitive aspects of EF might be a potential target for interventions that could have a substantial influence on everyday life for children and adolescents with ASD.

5.1.3 Sex differences in the association between EF and social difficulties In all three papers, we found that the sex of a participant influenced the results, which supports the hypothesis that different mechanisms may be at play in girls and boys with ASD. Another possible explanation is that girls and boys have different clinical manifestations of ASD. The most striking sex difference was that girls have a stronger relationship between EF deficits and autistic symptomatology related to social interaction and communication (BRIEF and ADI-R scores) than boys (Paper II). We also found girls to have a stronger association between EF and a questionnaire that defines social function (SRS), relatively broadly (Paper I). However, differences in correlation coefficients between girls and boys were not significant after correction for multiple comparisons in the last-mentioned findings. The results support the assertion that there may be different reasons for cognitive and behavioral problems in girls and boys with ASD, and that social and communications challenges for girls might be more strongly related to EF deficits. In addition, the results may to some extent be regarded as support for the female protective effect, which states that girls have a higher genetic threshold for developing ASD. A possible consequence of this higher threshold for developing ASD for girls is that girls who reach the diagnostic threshold of ASD often have more cognitive and behavioral problems than boys with ASD (Dworzynski et al., 2012). Furthermore, the girls in our studies showed a tendency to be older on average than the boys, had more difficulties with metacognitive aspects of EF and social function (SRS), and had less difficulties relating to restricted and repetitive behaviors. These findings are in line with previous research showing that clinical populations of girls with ASD usually are identified later, have more cognitive and behavioral problems, and exhibit less problems relating to

49 restricted and repetitive behaviors than boys with ASD (Beggiato et al., 2017; Dworzynski et al., 2012; E. I. White et al., 2017).

The extent to which the results are sensitive to sex differences may depend on the sensitivity of the assessment method used. The SRS questionnaire has sex/gender specific norms and is thought to better capture the female autistic phenotype than other measures (Beggiato et al., 2017). On the other hand, the diagnostic interview ADI-R has been criticized for being more sensitive to typical boys’ autistic features than their female counterparts. Beggiato and colleagues argue there is a need for diagnostic algorithms differentiated by sex in ADI-R, and that ASD symptoms in girls will be under-recognized if diagnostic instruments do not reflect the sex differences in clinical presentations (Beggiato et al., 2017). They emphasize that girls show less problems relating to the items in the algorithm concerning repetitive and stereotyped behaviors, which is a consistent finding in the literature (Frazier, Georgiades, Bishop, & Hardan, 2014). However, even though the girls in our sample showed a tendency for less restricted and repetitive behaviors than boys on the ADI-R, it is the relationship between EF deficits and social dysfunction that is our focus, and not the cut-off limits for diagnosing ASD. As earlier mentioned, Chouinard et al. also found sex differences to differentially affect the relationship between EF and social communication in children and adolescents with ASD (Chouinard et al., 2019).

In Paper III, we found the sex of group members to have a significant influence on the regression analysis results investigating the relationship between PGS group and behavior regulation aspects of EF (BRIEF). However, we did not investigate sex differences with further analyses, and cannot draw any firm conclusions about the finding. Some evidence suggests that girls have more serious genetic conditions (CNVs) and that common variants, as measured by PGS, are more important for boys (Werling, Parikshak, & Geschwind, 2016). We did control for CNVs in our sample, and it is therefore not likely that differences in CNVs were the reason for differences in the low and high PGS groups. Both Benca et al. (2017) and Schork et al. (2018) incorporated the sex variable as a covariate in the analysis when studying the association between PGS and EF (Benca et al., 2017; Schork et al., 2018). However, they do not comment on the effects of the sex variable on the relationship between PGS and EF in their discussion of their findings.

50 To summarize, we found that the girls in our studies were, on average, older than the boys. The girls had more metacognitive EF difficulties and social difficulties measured by the SRS. The boys in our sample exhibited more problems than the girls relating to restrictive and repetitive behaviors. However, these differences between girls and boys were not significant after correcting for multiple comparisons. We found no sex differences in PGS scores or the distribution of comorbid ADHD. Therefore, both girls and boys were quite similar in the measures used in this thesis, even though it could be argued that some measurements are more sensitive to difficulties experienced by boys than girls. Our main finding is that girls had a different relationship between social dysfunction and EF than boys, where the girls showed a stronger association between the two. This might be important for understanding the different manifestations of ASD in boys and girls and may have implications for choice of intervention.

5.1.4 Genetic relationship to EF The main reason for using PGS was to try to disentangle psychological constructs like EF, intelligence and social function. In particular, we wanted to use a biological measurement like PGS and investigate how this is associated to behaviorally defined diagnoses and psychological concepts like executive functioning. PGS is a relatively new method of assessing genetic information and calculating genetic risk, and even though PGS still only explains a few percentiles of the variance for a disease/trait in a given population, it might add valuable information as sample sizes for the traits and disorders increase. At the present time, PGS cannot be used diagnostically. A high PGS for ASD does not necessarily mean that a person will develop the disorder, and a low score does not mean that they will not. PGS cannot replace screening either. However, it is possible that PGS in the future can be used to assess risk populations and tailor diagnostics and interventions, cf. precision medicine (Zheutling & Ross, 2018).

Grove et al. demonstrated that an individual’s risk for ASD depends on the level of polygenetic burden of common genetic variants (Grove et al., 2019). In Paper III, where we used PGS, we found a significant difference in ASD PGS between individuals with ASD and a small group of healthy controls (n=29). We also found a significant difference between the controls and clinical participants consisting of children and adolescents with ASD symptomatology. However, the ASD PGS did not significantly differ between those with ASD and the clinical group with ASD symptoms (sub-threshold ASD group). In child and adolescence psychiatry, it can be difficult to accurately diagnose people because of

51 overlapping clinical characteristic, common comorbidities and the lack of objective measures and biomarkers. Furthermore, the clinical features might change during development from childhood and adolescence to adulthood. Assuming that autistic traits and ASD PGS are present in everyone on a continuum, it is expected that children in a clinical setting under assessment for ASD have a higher PGS for ASD than TD children. Currently, PGS does not solve the demanding task of distinguishing between different related diagnoses. However, it might be particularly interesting to look at “extreme” PGS scores, as some have done (Torkamani et al., 2018). To explore this, we divided our participants into low, moderate and high scores based on their PGS. In Paper III, we found approximately the same distribution of children with an ASD diagnosis in both the low and high ASD PGS groups. Interestingly, the relationship to behavior regulation aspects of EF was different in the low and high ASD PGS groups, where the high group had more difficulties with EF. Our finding supports the link between the higher risk and higher rate of EF difficulties seen in clinical populations. The behavior regulation aspect of the EF measure we used consists of the three subscales inhibition, flexibility and emotional control. Both in our sample and in previous research, the flexibility subscale is where people with ASD most often have the highest score (i.e. most difficulty) compared with other clinical groups (Hovik et al., 2014). In this way, the aspect of EF that is usually most aberrant in people with ASD was most strongly associated with the common genetic variants related to ASD.

The use of extreme PGS scores seems promising and might together with other clinical information add valuable information, and PGS can be particularly helpful in targeting high- risk people (Torkamani et al., 2018). Furthermore, people with high risk for a given disorder might also be at increased risk for known comorbid traits/ disorders and related difficulties, such as EF deficits. When researchers have access to PGS for several related disorders and traits, it might be possible to disentangle both mechanisms and characteristics of the conditions, such as looking into the interplay between PGSs for ASD, ADHD, general intelligence, and depression. PGS may in the future provide a valuable window into the core of neurodevelopmental disorders and allow us to look behind the behaviorally defined diagnoses.

We did not find a significant contribution of the PGSs for ADHD or general intelligence (INT) to EF in our study. Thus, the common genetic variants associated with ASD seem to be of greater importance for executive dysfunction than the genetic variants associated with

52 ADHD and INT. One possible explanation for this may be that our sample consisted of children who were undergoing assessment for ASD, and that the difficulties they showed were primarily related to ASD symptoms. Another interesting result was that we found a significant association between social dysfunction (high scores on the SRS) and high PGS for general intelligence (INT). This relationship amounted to a moderate effect size in the comparison of the low and high INT PGS and SRS score (Cohens d = 0.66). The result here is in line with an increasing number of studies linking genetic variants associated with high IQ to social problems (Clarke et al., 2016; Grove et al., 2019).

Only a few studies have investigated the association between PGS for ASD and EF. Schork et al. examined the relationship between PGS for ASD and EF in TD children. They used a card sorting test to measure EF and found that better performance on this test was associated with high PGS (Schork et al., 2018). This result is the opposite direction of what might be expected, as ASD is usually associated with executive dysfunction (Demetriou et al., 2017). However, it is important to note that Schork et al. investigated TD children. A possible explanation for the finding may be that the card sorting test they used is related to IQ, and high ASD PGS has been found to be associated with high IQ (Clarke et al., 2016; Roca et al., 2014). Benca et al. used a population-based sample to investigate if PGS for psychiatric disorders, including ASD and ADHD, could predict EF (Benca et al., 2017). They used a latent variable approach based on nine EF tasks, but did not find evidence of a strong connection between EF and PGS in psychiatric disorders. Our study differs from these studies since we used a clinical sample and a parent-rated measure of EF. However, our finding is in line with earlier findings that EF deficits are common in ASD, and that parent-rated measures are a better clinical marker of ASD than neuropsychological tests (Demetriou et al., 2017).

There is an interesting association between high IQ and higher PGS for ASD in the general population (Clarke et al., 2016). Furthermore, Grove et al. found that people with Asperger had higher PGS than those with classic childhood autism (Grove et al., 2019). By definition, people with Asperger have normal to above normal intelligence, and in this way higher IQ seems to be related to higher PGS in the ASD population. It seems likely that PGS may have different effects depending on the specific ASD subgroup. Since our sample mainly consisted of people diagnosed with Asperger or PDD-NOS, we used the ASD PGS on the subgroups where it has been shown to have the greatest effect.

53 In addition to common genetic variants with little impact, ASD is also associated with several rare genetic disorders with high impact (Levy et al., 2011). CNVs are probably the strongest factor contributing to these rare genetic disorders. We did have access to information about the most common recurrent CNVs and other mutations in our participants. Based on this information, it is not likely that the identified association between high PGS and EF deficits was due to the presence of CNVs or other severe genetic causes.

One of the main purposes of PGS has always been precision medicine (Zheutling & Ross, 2018). US National Institute of Health defines precision medicine as “an emerging approach for disease treatment and prevention that takes into account individual variability in genes, environment, and lifestyle for each person” (National Insitute of Health, 2020). This approach contrasts the one-size-fits-all approach, in which disease treatment and prevention strategies are developed for the average person, where less attention is paid to individuals’ specific needs. By using precision medicine, it will be possible for health providers and researchers to predict more accurately which treatment and prevention strategies for a particular disease will work in which groups of people. Because of the large variation in symptoms and function in people with ASD, precision medicine might be particularly useful for this patient group. Although our results regarding the association between high PGS for ASD and EF difficulties must be interpreted with caution, it may be possible that PGS in the future could be used to tailor treatment and interventions for people with ASD.

5.1.5 The association to comorbid ADHD Since ADHD is a common comorbidity in ASD and is also characterized by EF deficits, we considered it important to include participants with comorbid ADHD in all our papers. In our studies, 33-35% of the participants had a comorbid ADHD diagnosis (Paper I-III). This is in line with previous studies, which report 28-44% comorbid ADHD in ASD (Lai et al., 2014; Lord et al., 2018; Simonoff et al., 2008). We found no significant differences in social difficulties (total SRS), full-scale IQ or behavior regulation EF (BRI form BRIEF) between those with ASD and those with ASD and comorbid ADHD (Paper III). We did find, however, that those with ASD and ADHD had significantly more difficulties with metacognitive aspects of EF (MI from the BRIEF) and a higher overall score on the BRIEF (GEC) than those who only had ASD. This might suggest that children and adolescents with comorbid ADHD have more EF deficits than children diagnosed with ASD alone. Unfortunately, we were not able to analyze in more detail these EF differences. Our findings are in line with

54 Craig et al.’s conclusion in their review of executive dysfunction in ASD and ADHD. Craig et al. found that people with ASD with comorbid ADHD have EF deficits that are typical of both the ASD and the ADHD group, and they therefore often have more EF deficits than those who only have one of the diagnoses (Craig et al., 2016).

We incorporated ADHD into the analyses in all three papers. In Paper I, we found that the relationship between social difficulties (SRS) and metacognitive aspects of EF (MI from the BRIEF) was stronger when we excluded the participants with comorbid ADHD from the analyses. However, the metacognitive aspects of EF were significantly related to social difficulties, both for the participants with ASD and the participants with ASD and comorbid ADHD.

In Paper II, where we investigated possible sex differences in the relationship between autism symptomatology and EF, we found no significant difference in the proportion of males and females with comorbid ADHD. It is therefore not likely that the presence of ADHD is the explanation for the difference between the sexes. Furthermore, ADHD did not represent a significant contribution in any of the regression models predicting social reciprocity, communication or restricted and repetitive behaviors and interests. Overall, we did not find that ADHD significantly affected the relationship between autism symptomatology and EF in our study.

When investigating the relationship between PGS for ASD and EF, we divided the groups into those with low, medium and high ASD PGS. We did not find a significant difference in the distribution of ADHD or ASD diagnoses in the low and high ASD PGS groups. It is therefore not likely that low and high PGS groups were purely a reflection of diagnostic groups. In Paper III, we also incorporated the ADHD PGS in the prediction of EF deficits and social difficulties. However, our results did not show that the ADHD PGS group contributed significantly to the EF (BRIEF) or social function (SRS) scores.

In sum, our results suggest that those with comorbid ADHD had more difficulties with the metacognitive aspects of EF than those who only had ASD. However, the relationship between social function and EF is not significantly different between those with ASD only and those with comorbid ADHD. The effect of comorbid ADHD may vary with age. Since we used a cross-sectional design, we did not investigate how the influence of age may change

55 over time. A diagnosis of comorbid ADHD can have important clinical implications, because some ADHD symptoms can be remedied with medication (Mooney et al., 2019). The various diagnostic systems have taken different approaches as to whether it is possible to have a comorbid ADHD diagnosis for people with ASD. Based on our results, comorbid ADHD was seen to influence the clinical manifestation of difficulties to some degree. However, we did not find that comorbid ADHD significantly affected the relationship between EF and social difficulties.

5.1.6 Intelligence and EF Many researchers argue that EF is closely related to intelligence, and particularly to fluid intelligence (Diamond, 2013; Roca et al., 2014). Other researchers maintain that people with ASD, for example, have EF impairments despite having normal or above normal levels of IQ (Merchan-Naranjo et al., 2016). To account for IQ potentially affecting the relationship between EF and social function, we incorporated full-scale IQ into all our analyses. IQ did not significantly influence the results in any of the three studies. One explanation may be that our sample was small and therefore lacked power. Another explanation may be the large variation in IQ scores. Given that many researchers find IQ to be related to EF, our finding is to some extend unexpected. However, one must take into account that we used a parent-rated measure of EF and not neuropsychological test results. There might have been a closer connection between laboratory performance measures of EF and intelligence. On the other hand, the relationship between intelligence, EF and social function may be different in people with ASD compared to other neurodevelopmental disorders or TD children (Tillmann et al., 2019). Despite ASD being genetically associated with high IQ, people with ASD have profound difficulties with social function and EF (Clarke et al., 2016).

5.1.7 General considerations about the association between executive dysfunction and social difficulties in ASD The current studies address the knowledge gap in the literature by investigating the relationship between a core autistic difficulty (social function) and executive dysfunction in everyday life in children and adolescents with ASD or ASD symptomatology. ASD is behaviorally defined because there is limited knowledge about mechanisms and few medical treatments for autistic symptoms. Identification, diagnosis and interventions relating to ASD are based on characteristics of core autistic features and accompanying cognitive difficulties.

56 The treatments that have most empirical evidence of efficacy are behavioral, developmental and/or educational. Cognitive aspects also influence these treatments (Odom et al., 2019). General cognitive abilities are often affected in ASD, and language difficulties and EF deficits are common. Furthermore, co-occurring psychiatric disorders such as ADHD (also associated with EF deficits) are often present. Therefore, understanding how social difficulties in ASD relate to EF can add valuable knowledge to the field of ASD.

One of the prevalent theories of ASD has been the Executive Dysfunction Hypothesis. Much of the research on this theory has focused on examining to what extent people with ASD have EF deficits, and if this can be used as part of the diagnostic criteria. Reviews and meta- analyses support the view that difficulties with EF are common in people with ASD (Demetriou et al., 2017). However, the most important criticism of the theory is probably related to EF deficits not being specific to ASD. Identifying the exact type of EF deficits in people with ASD has also been difficult, and EF difficulties on performance based tests do not necessarily correlate with autistic symptoms (Kenworthy et al., 2008; Pellicano, 2010). Furthermore, there has been a lack of research investigating how social difficulties/autistic symptoms and EF affect each other. By focusing on the relationship between them rather than just the presence of EF difficulties in ASD, we might uncover knowledge that can help in understanding disease mechanisms and find potential intervention approaches. Since ASD is a very heterogeneous diagnostic group, looking at them as a single group might mask important differences. Therefore, we wanted to investigate different subgroups of the ASD population, and explore how variables such as the sex of the individual (boy/girl), comorbid ADHD and high PGS for ASD affect the relationship between social function and EF.

In all three papers, we found a relationship between EF and social function and/or autism symptomatology in ASD. Furthermore, methods such as PGS, diagnostic tools and parent- reported difficulties in everyday life were used to investigate this relationship. We found that the association between EF and social function was valid on different levels of explanation – from genes and cognition to autistic behavior.

57 5.2 Methodological issues 5.2.1 Representativeness and generalizability of results The participants in our studies were recruited from Norwegian health services specializing in the assessment of ASD and neurodevelopmental disorders. I regard the participants to be representative of children and adolescents who seek clinical help for ASD in clinics throughout Norway. However, those who seek help might not necessarily be representative of the population of children with ASD and might have a higher symptom load than those who do not seek help. Furthermore, we do not have any information about those who did not want to participate, and if they had any common characteristics that made them not want to participate in the study. In Paper I and Paper II, we only included participants who received a diagnosis of ASD. In Paper III, we also included participants with sub-threshold ASD/ ASD symptoms, which is the inclusion criteria for the BUPGEN project. Therefore, the finding regarding the association between ASD PGS and behavior regulation aspects of EF is probably representative of a broader group with ASD symptomatology, who do not reach the diagnostic threshold. Since ASD is a behaviorally defined spectrum disorder, features of autism-like traits are thought to vary throughout the population in a continuous way. Investigating a broader autism phenotype may add important knowledge to the field (Constantino, 2011). Genetically common variants, such as those measured with PGS, have also shown greater effect for those with an Asperger diagnosis than for those with classic autism. Common genetic variants may also be important for the sub-threshold ASD group (Grove et al., 2019). The broader inclusion that we have conducted may also be an advantage considering that the diagnostic tools probably are better at recognizing boys with ASD (the male phenotype) than girls with ASD. By broadening the inclusion, I can assess girls who might not otherwise have been included in clinical studies.

In all our papers, we only included people with a full-scale IQ of 70 and above. Our findings are therefore not representative of children and adolescents with an IQ under 70 and/or an intellectual disability. Since our participants are diagnosed using ICD-10, they are all grouped into the subcategories childhood autism, atypical autism, Asperger syndrome or PDD-NOS. A natural consequence of having a selection of participants with an IQ above 70, is that they mainly fall into the diagnostic categories Asperger syndrome and PDD-NOS, which is also the case in our sample (these two categories constitute 79-84% of the participants with ASD in our studies). Our findings might therefore mostly be applicable to these two diagnostic

58 subcategories. Furthermore, we included participants with comorbid ADHD, which is the most common comorbidity in ASD (Lord et al., 2018). Our results may therefore be generalized to children and adolescents with ASD with an IQ above 70 and comorbid ADHD.

We had a fair number of girls in all three studies, as girls with ASD constituted 22-27% of the participants. Even though this is quite close to the ratio 4:1, which is often reported in the literature, the ratio is often more skewed in participant with IQ in the normal range with even more boys than girls (reported to be as high as 9:1) (Brugha et al., 2011). We have quite a high number of girls in our studies, which might be due to a recent increased awareness of girls with ASD and their symptom profiles. A large part of the research literature mainly consists of boys and is not necessarily applicable to girls with ASD. However, we do consider our findings to be valid for girls, even though the actual number of girls included is still relatively low (23-42 girls with ASD in our studies).

One important criticism of the use of PGS is that most of the calculations are based on European samples, and much of the genomic data comes from the Scandinavian and UK biobank. The calculations may therefore not be representative of other ethnic groups, and the generalizability of the findings may be limited to mainly Northern European samples (Martin et al., 2019).

5.2.2 Possible confounding factors All three of our studies are based on clinical samples, and not population-based samples or research samples. As a result, the participants included are patients who seek help and might therefore have a higher symptom load than those who do not seek help. Parents might also report symptoms and behaviors differently for girls and for boys, and it is uncertain if this is due to different behaviors/ symptoms in general between the sexes or whether parents have different expectations of girls and boys. In this thesis, we use the term sex differences. However, we do not know if the differences are caused by actual biological differences or differences relating to expectations of gender and gender role patterns. Both SRS and BRIEF have specific norms for girls and boys, and thereby take into account to what extent they have difficulties compared to TD girls and boys. Furthermore, children and adolescents undergo major developmental changes in social function, IQ, and EF as they grow older. Girls and boys might also have different trajectories of development (Grissom & Reyes, 2019). Even

59 though we incorporated age as a covariate in our analyses, non-linear age effects could bias the results.

Our participants were assessed cognitively by using intelligence measures, but we did not perform a comprehensive assessment of language functions. Difficulties with language and communication can affect both social function and EF in people with ASD, even though there is variability in how these difficulties are manifested (Lord et al., 2018). Some studies have found that metacognitive aspects of EF are predictive of functional communication, while behavioral aspects of EF were predictive of verbal conversation skills (Hutchison, Muller, & Iarocci, 2019). Others have shown that EF level was not related to language ability in verbal, school-aged children with ASD (Joseph, McGrath, & Tager-Flusberg, 2005). Our sample consisted of school-aged children and adolescents with normal IQ.

Since we included participants with comorbid ADHD, some of the children and adolescents may have been taking medication that is known to have a positive impact on neurocognition (Coghill et al., 2014). When answering the questionnaires (SRS and BRIEF), the parents are asked to report behavior the past 6 months. We only have information about medication at one given time point and cannot be sure if they had used medication during the previous 6 months. However, most of our participants were recruited at the start of an assessment period for a possible neurodevelopmental disorder, and therefore likely not to be on medication at the time of inclusion.

5.2.3 Measurements Participants underwent a thorough clinical evaluation for ASD using the gold standard tools, ADOS and/or ADI-R, administered by an experienced clinician (child psychologist and/or child psychiatrist) (Lord et al., 2000; Lord et al., 1994). Since participants were recruited from a clinical setting, the diagnostic assessment and final conclusion were also based on more general information about the child’s mental and somatic health. Information was obtained from anamnestic interviews, somatic evaluations, questionnaires measuring general psychopathology, clinical interviews, naturalistic observations and formal testing of cognitive function. We are confident that the ASD diagnoses were arrived at in a professional manner. Participants were recruited from different BUPGEN locations throughout Norway. To ensure that we had a common diagnostic practice, we held diagnostic workshops across the different clinics. Furthermore, there are guidelines for diagnosing and treating ASD in Norway. A

60 recent study confirmed that ASD assessments in Norway are conducted in accordance with the guidelines, and in 95% of the cases examined, the diagnostic research criteria had been satisfied (Suren et al., 2019). We cannot rule out, however, that there may be some degree of variation in diagnostic practices used in our sample.

To assess everyday life EF, we used the questionnaire BRIEF. In Paper I and Paper III, we used SRS to measure social impairments associated with ASD. Both BRIEF and SRS are well-validated instruments/questionnaires commonly used in both research and clinical practice (Constantino & Gruber, 2005; Gioia et al., 2000). The average t-score for TD Norwegians in BRIEF and SRS has been found to be about 40, while the average t-score in American norms is 50 (Hoyland, Naerland, Engstrom, Lydersen, & Andreassen, 2017; Skogli, Teicher, Andersen, Hovik, & Oie, 2013). Yet, the BRIEF indexes are regarded to work well when used for the Norwegian population (Fallmyr & Egeland, 2011). Our sample is a clinical sample and on average both the BRIEF and SRS scores in our studies were in the clinical range. Since both BRIEF and SRS are parent-rated measures, this might bias the findings. Parents can have a tendency to systematically over- or under-rate the degree of disturbance in their child. It can be difficult to determine if high scores on two questionnaires are due to a common underlying trait (i.e. trait variance) or due to the use of a common method (i.e. method variance) (McAuley, Chen, Goos, Schachar, & Crosbie, 2010). The assessment is also influenced by the raters’ theoretical and logical preconceptions. Havdahl et al. found that parental concern about ASD significantly influenced the discriminative threshold of the ADI- R, and the instrument had lower sensitivity among toddlers whose parents did not have a specific concern about ASD (Havdahl et al., 2017). On the other hand, an advantage of using measures such as SRS and BRIEF is that they are likely to have high ecological validity, which means that they are good at predicting behaviors in real-world settings (Constantino & Gruber, 2005; Gioia et al., 2000). For example, people with ASD often show rigid and inflexible everyday behaviors, even though they do not necessarily have deficits with flexibility on laboratory measures (Geurts, Corbett, & Solomon, 2009). This discrepancy is referred to as the paradox of assessing cognitive flexibility in ASD, and emphasizes the importance of having ecologically valid measurements. A similar pattern is seen between IQ measures and the Vineland adaptive scale, where people with ASD often have a greater discrepancy between the two measurements than other groups/ people (Tillmann et al., 2019).

61 The main focus of this thesis has been on social function and its relationship to EF. An ASD diagnosis is based on deficits in social communication and social interaction across multiple contexts, and restricted, repetitive patterns of behaviors, interests, or activities (American Psychiatric Association, 2013). Our focus has been on issues relating to social communication and social interaction, which together is what we refer to as social function. To some extent, we have also investigated the relationship between EF and restricted and repetitive behaviors and interests, even though this has not been our primary research focus. In SRS, only one of the five subscales is related to restricted and repetitive behaviors and interests, namely the autistic mannerisms subscale (which is incorrectly called social mannerisms in Paper I). Furthermore, in ADI-R, one of the three functional domains (domain C) is related to restricted and repetitive behaviors and interests. If we wanted to investigate restricted and repetitive behaviors and interests and its relationship to EF in further detail, instruments more suitable for capturing difficulties with restricted and repetitive behaviors and interests should be used (Honey, Rodgers, & McConachie, 2012).

Choi et al. provide guidelines on how to perform PGS analysis, which has the advantage of limiting inconsistencies between study protocols and misinterpretation of results. Even though our scores are standardized (normal distribution), there is no general agreement on standardization. At the present time, it is therefore difficult to compare the actual value of one score with another (Choi, Mak, & Reilly, 2018). The PGSs used in our study were based on large discovery samples with large power and performed by high quality research groups. The discovery samples consisted of 18,381 ASD cases and 27,969 controls for the ASD PGS (Grove et al., 2019), 20,183 ADHD cases and 35,191 controls for the ADHD PGS (Demontis et al., 2018), and 269,867 cases for the INT PGS (Savage et al., 2018). We therefore consider our PGSs to be of good quality. When using PGS, it is common to do the calculations at several different p-levels (p-bins). In Paper III we used a GWAS p-level threshold for inclusion of 0.1 for all the PGSs (ASD, ADHD and INT). This threshold was chosen because it gave the highest explained variance in the GWAS summary statistics, which was the basis for our ASD PGS (Grove et al., 2019). Furthermore, it is important to be aware that some of the criticism of PGS in general is that they have a relatively low explained variance. However, the prediction will become more feasible as sample sizes continue to grow (Dudbridge, 2013). In Grove et al., where they calculated the ASD PGS we used, the mean variance explained by PGSs was 2.45% (Nagelkerke’s R2). The variance explained increased for each PGS included, and was 3.77% for the full model (Nagelkere’s R2) (Grove et al.,

62 2019). Values between 1-9% are usually interpreted as representing a small effect size (Cohen, 1992).

5.2.4 Questionnaires versus neuropsychological testing In a 2013 review, Toplak et al. argues that performance-based and rating measures of EF assess different levels of cognition. While the former measures the efficiency of cognitive abilities, the latter measures success in goal pursuit (Toplak et al., 2013). Performance-based measures attempt to isolate a single structure and might be a more “pure” measurement of EF, while parent-ratings may represent a more global view of the child’s behavior in everyday life (Loe, Chatav, & Alduncin, 2015). Research has shown only moderate correlations between performance-based measures and questionnaires assessing EF (Isquith, Roth, & Gioia, 2013; Toplak et al., 2013). Furthermore, it might be particularly challenging to measure EF in ASD populations because they can perform quite well in structured situations like a formal test situation, even though they might have substantial difficulties in everyday life (Geurts et al., 2009). Most studies linking EF to brain functions are based on neuropsychological test results. Associations have been found between structural MRI results and ratings of everyday EF as well (Isquith et al., 2013). Given the moderate connection between test performance and behavior ratings, there is reason to believe that our finding would be different if we used neuropsychological tests to measure EF. For example, Landa and Goldberg only found a weak to non-existent association between neuropsychological tests of EF and social performance (Landa & Goldberg, 2005). On the other hand, it is likely that the metacognitive aspects of BRIEF are an expression of cognition, although what we primary measure is behavior. The behavioral aspects of BRIEF are naturally more closely related to behavior and emotional control. Taken together, it is a common understanding that both neuropsychological tests and questionnaires are of value, and that the two methods complement each other. However, it is likely that the association between EF and social function might differ depending on the measure used to assess EF. In a recent study of executive attention in children with ADHD, the researchers found, contrary to their prediction, that performance-measures did not explain unique variance of ADHD problems beyond that explained by behavior ratings (Tiego, Bellgrove, Whittle, Pantelis, & Testa, 2019). They therefore question the utility of performance-based measures of self-regulation in neurodevelopmental samples.

63 5.3 Implications 5.3.1 Theoretical implications Our studies confirm that there is a connection between EF deficits in everyday life and core autistic features like social difficulties in children and adolescents with ASD. Our findings can thus be viewed as support for the Executive Dysfunction Hypothesis of ASD (Hill, 2004; Pennington & Ozonoff, 1996). The aim of our studies was not to investigate the specificity and sensitivity of the Executive Dysfunction Hypothesis of ASD or other theories of ASD; nor did we aim to find explanations that applied to the entire ASD group. Our aim was to investigate the variance in the relationship between EF and social function in different subgroups, such as girls/ boys and those with comorbid ADHD. Since we used a parent-rated assessment of EF, we did not specifically measure the three EF subdomains (common EF/inhibition, flexibility and working memory) that constitute Miyake’s definition of EF (Miyake et al., 2000). However, we did find that some aspects of EF were more important than others. When it comes to social function in everyday life, metacognitive aspects of EF were particularly important, and difficulties with behavior regulation were associated with high PGS for ASD. Thus, we found support for the hypothesis that specific aspects of EF can have different relationships to risk factors and challenges associated with an ASD diagnosis. We also confirmed possible sex differences in ASD, where we found that even though boys and girls did not significantly differ in levels of social difficulties, EF deficits or PGS in our studies, boys and girls exhibited different relationships between social dysfunction and EF deficits, and between PGS and EF. Our finding that high PGS is related to behavior regulation aspects of EF is also one of the first times an association has been found between PGS and EF deficits measured in everyday life in children and adolescents with ASD symptomatology.

5.3.2 Clinical implications One of the main reasons for investigating the relationship between EF and social difficulties was to better understand the patient group and their needs, and to use this knowledge to inform clinicians about potential clinically relevant results.

The finding that metacognitive aspects of EF had a stronger connection to social dysfunction than behavior regulation aspects of EF is particularly important to be aware of in the clinic, because people with ASD are primarily characterized by problems relating to flexibility.

64 Therefore, it is possible that clinicians do not pay enough attention to metacognitive aspects of EF, and its close relationship to social difficulties.

Our findings also suggest that there may be a stronger association between EF deficits and social difficulties in girls than in boys. Therefore, it is important to be aware of EF deficits in girls under assessment for ASD, and to consider possible social challenges in girls under clinical assessment who present with EF problems. Furthermore, the association between EF deficits and ASD symptoms in girls seems to be stronger for the social interaction and social communications domains than for restricted and repetitive behaviors and interests. Based on our finding, it seems that restricted and repetitive behaviors and interests have a weaker link to EF. As a consequence, EF treatments may not significantly affect this area of ASD.

The current results could form the basis of developing more efficient and customized treatment. Many treatments for ASD target cognitive function, even though there are still relatively few interventions that specifically target EF. One reason why this area seems particularly promising is because it looks as if EF improvements can lead to better social function. Treatments targeting social difficulties mainly have an effect on social function and not EF difficulties (Kenworthy et al., 2014). Specific EF interventions are particularly interesting if the effects are generalizable to core ASD symptoms such as social difficulties. Although there are inconsistent findings regarding executive dysfunctions in people with ASD, a recent longitudinal study confirms that EF deficits in early childhood for a group of people with ASD has prognostic significance for autistic features and adaptive function 12 years later (Kenny, Cribb, & Pellicano, 2019). This underlines the importance of EF as a target for early intervention and support in ASD.

Since ASD is a homogeneous group, an important follow-up question is to ask which subgroups would most likely benefit from EF interventions. In our studies, we found a stronger relationship between social difficulties and EF in girls than in boys. Therefore, targeting EF deficits is probably more important for girls than boys. However, our study is a cross-sectional study, and we cannot know the directions of this relationship or the cause of the association. Surprisingly few studies discuss whether there are sex differences in treatment outcomes after EF or ASD interventions. In a review and meta-analysis of cognitive interventions for children with neurodevelopmental disorders, acquired brain injuries or neurological disorders, they do not mention how the effect of the interventions might be

65 influenced by the sex of the participants (Robinson, Kaizar, Catroppa, Godfrey, & Yeates, 2014). Based on our findings, we recommend being aware of possible sex differences when deciding treatment/intervention choices and when considering the efficacy of treatment. In clinical practice today, we do not offer any sex-specific treatments.

PGS methods have no clinical relevance for ASD diagnostics or treatment in child and adolescent psychiatry today. However, it can be useful for predicting genetic risk for a given disorder or trait and can help uncover a causal relationship between trait and condition. Even though PGS approaches are not used in the clinic currently, PGS may become a useful tool in diagnostics and for stratifying interventions in the future. In particular, those with a high PGS for ASD might be at risk for both the disorder and additional difficulties such as EF deficits. In our sample, participants in the low ASD PGS group had on average a t-score of 62.9 on the BRI, which is under the clinical borderline/threshold. On the other hand, participants in the high ASD PGS groups had an average t-score of 71.5 on the BRI, which is well within the clinical range. This difference in t-scores represents a medium effect (Cohen’s effect size of d = 0.69).

When looking at other conditions and diseases, an approach that combines PGS with other types of information can significantly increase the odds ratio for a given disease and/or treatment outcome. For example, PRS has been used to stratify breast cancer risk and to explore implications for screening (Lewis & Vassos, 2017). Andersson et al. have conducted a very interesting study where they explored how PGS for several psychiatric disorders were related to treatment response to cognitive behavior therapy (CBT) for depression (Andersson et al., 2018). They found high ASD PGS to be associated with less effect of CBT. Anderson et al.’s study is the first to find an effect of PGS on treatment outcomes in psychiatric disorders and this might be an interesting approach to apply in the future for similar applications. However, replication of the results is needed before any firm conclusions can be made. Uddin et al. take this even further and offer a description of how artificial intelligence and precision medicine can be used in neurodevelopmental disorders in the future. They recommend stepping away from existing diagnostic constructs, and look at biologically driven phenotypes to see how these maps onto treatment response and clinical outcomes (Uddin, Wang, & Woodbury-Smith, 2019).

66 5.4 Strengths and limitations, and future research A strength of the current study is that it includes a reasonable sample size of clinically well- defined children and adolescents with ASD or ASD symptomology using gold standard measures like ADOS and ADI-R. Another strength is that we have a moderate number of girls in our cohort. A third strength is that we had access to a biological measurement such as PGS, and that all the PGSs were based on large discovery samples. BRIEF and SRS are on the other hand measurements related to everyday function, with less control over specific mechanisms, but with high ecological validity. By using measures related to everyday function it might be easier to transfer the knowledge to clinical practice and everyday life of the participants.

The age range and total IQ range in the samples are quite large, and this might represent a limitation. However, we incorporated age and total IQ into the analyses in all three papers to account for the impact these variables may have on the results. All of the comorbid ADHD diagnoses were based on clinical assessment and ICD-10 criteria. However, we did not have a measurement of ADHD symptom severity. Furthermore, we only had information from parents and not teachers regarding everyday functioning (BRIEF and SRS). We did not have access to both ADOS and ADI-R/SCQ performed in the same time period for all the participants. Even though the PGSs are based on large control groups and the questionnaires (BRIEF and SRS) have normative data, we did not have a control group (with the exception of the 29 controls in Paper III). Lastly, the study had a cross-sectional design, and we can therefore not make any conclusions regarding developmental trajectories or causal relationships between EF deficits and social difficulties.

Future research should further elaborate on the relationship between EF and autistic symptomatology in more detail. In particular, it would be of interest to explore whether there are specific subgroups of persons with ASD (e.g. age, sex, IQ and comorbid conditions) with specific needs. Furthermore, it would be interesting to investigate EF deficits by using neuropsychological tests of both hot and cold EF in addition to parent-rated measures. Integrating genetic aspects by using PGS in combination with cognitive and clinical information also represents an exciting opportunity in the future. Lastly, children and adolescents are constantly developing, and it would be particularly interesting to study developmental trajectories in both EF and social function through a longitudinal research design.

67 6. Conclusion Based on the studies described in this thesis, we found support for a relationship between parent-rated EF in everyday life and social dysfunction in children and adolescents with ASD or ASD symptomatology.

We found the metacognitive aspects of EF to be particularly important for social function in everyday life. Furthermore, we found a tendency for this association to be stronger for girls and for the older age group (13-18 years). When we investigated the relationship between autism symptomatology based on the diagnostic tool ADI-R and parent-rated EF, we found the relationship to be significantly stronger for girls than for boys. This finding was related to social interaction and social communication aspects of ASD, and not to restricted and repetitive behaviors and interests. The presence of comorbid ADHD did not have a significant influence on the relationship. Lastly, we found high PGS for ASD to be associated with behavior regulation aspects of EF in a group of children and adolescents under assessment for ASD. This study is one of the first to find a connection between high PGS for ASD and EF deficits in everyday life.

Overall, these findings extend our knowledge in the field of ASD research by describing not only that children and adolescents with ASD have EF deficits, but by investigating how EF deficits and social difficulties are related to each other and which subgroups are particularly vulnerable to experience these difficulties. This knowledge can have important implications for understanding and treating ASD and related neurodevelopmental difficulties.

7. References Achenbach, T. M. (2011). Commentary: Definitely more than measurement error: but how should we understand and deal with informant discrepancies? J Clin Child Adolesc Psychol, 40(1), 80-86. doi:10.1080/15374416.2011.533416 Allan, C. M., Campbell, W. N., Guptill, C. A., Stephenson, F. F., & Campbell, K. E. (2006). A conceptual model for interprofessional education: the international classification of functioning, disability and health (ICF). J Interprof Care, 20(3), 235-245. doi:10.1080/13561820600718139 Alvarez, J. A., & Emory, E. (2006). Executive function and the frontal lobes: a meta-analytic review. Neuropsychol Rev, 16(1), 17-42. doi:10.1007/s11065-006-9002-x American Psychiatric Association. (2013). Diagnostic and Statistical Manual of Mental Disorders (DSM-5). Washington, DC: APA.

68 Andersen, P. N., Skogli, E. W., Hovik, K. T., Geurts, H., Egeland, J., & Oie, M. (2015). Working memory arrest in children with high-functioning autism compared to children with attention-deficit/hyperactivity disorder: results from a 2-year longitudinal study. Autism, 19(4), 443-450. doi:10.1177/1362361314524844 Andersen, P. N., Skogli, E. W., Hovik, K. T., & Øie, M. G. (2016). Development of executive functions over a two-year period in youth with high-functioning autism or ADHD. Nevropsykologi, 2. Andersson, E., Crowley, J. J., Lindefors, N., Ljotsson, B., Hedman-Lagerlof, E., Boberg, J., . . . Ruck, C. (2018). Genetics of response to cognitive behavior therapy in adults with major depression: a preliminary report. Mol Psychiatry. doi:10.1038/s41380-018- 0289-9 Asperger, H. (1944). Die Autistischen Psychopathen im Kindersalter. Arch Psychiatr Nevrvenkr(117), 76-136. Bai, D., Yip, B. H. K., Windham, G. C., Sourander, A., Francis, R., Yoffe, R., . . . Sandin, S. (2019). Association of Genetic and Environmental Factors With Autism in a 5- Country Cohort. JAMA Psychiatry. doi:10.1001/jamapsychiatry.2019.1411 Baio, J., Wiggins, L., & Christensen, D. L. e. a. (2018). Prevalence of Autism Spectrum Disorder Among Children Aged 8 Years. Autism and Developmental Disabilities Monitoring Network, 11 Sites, United States, 2014. MMWR Surveill Summ, 67(No. SS-6), 1-23. doi:http://dx.doi.org/10.15585/mmwr.ss6706a1 Bal, V. H., Kim, S. H., Fok, M., & Lord, C. (2018). Autism spectrum disorder symptoms from ages 2 to 19 years: Implications for diagnosing adolescents and young adults. Autism Res. doi:10.1002/aur.2004 Barendse, E. M., Hendriks, M. P. H., Thoonen, G., Aldenkamp, A. P., & Kessels, R. P. C. (2018). Social behaviour and social cognition in high-functioning adolescents with autism spectrum disorder (ASD): two sides of the same coin? Cogn Process, 19(4), 545-555. doi:10.1007/s10339-018-0866-5 Baron-Cohen, S. (2000). Is asperger syndrome/high-functioning autism necessarily a disability? Dev Psychopathol, 12(3), 489-500. Retrieved from https://www.ncbi.nlm.nih.gov/pubmed/11014749 Baron-Cohen, S. (2017). Editorial Perspective: Neurodiversity - a revolutionary concept for autism and psychiatry. J Child Psychol Psychiatry, 58(6), 744-747. doi:10.1111/jcpp.12703 Baron-Cohen, S., Leslie, A. M., & Frith, U. (1985). Does the autistic child have a «theory of mind»? Cognition(21), 37-46. Beggiato, A., Peyre, H., Maruani, A., Scheid, I., Rastam, M., Amsellem, F., . . . Delorme, R. (2017). Gender differences in autism spectrum disorders: Divergence among specific core symptoms. Autism Res, 10(4), 680-689. doi:10.1002/aur.1715 Benca, C. E., Derringer, J. L., Corley, R. P., Young, S. E., Keller, M. C., Hewitt, J. K., & Friedman, N. P. (2017). Predicting Cognitive Executive Functioning with Polygenic Risk Scores for Psychiatric Disorders. Behav Genet, 47(1), 11-24. doi:10.1007/s10519-016-9814-2 Bettelheim, B. (1967). The empty fortress: infantile auitsm and the birth of the self: Free Press. Bieleninik, L., Posserud, M. B., Geretsegger, M., Thompson, G., Elefant, C., & Gold, C. (2017). Tracing the temporal stability of autism spectrum diagnosis and severity as measured by the Autism Diagnostic Observation Schedule: A systematic review and meta-analysis. PLoS One, 12(9), e0183160. doi:10.1371/journal.pone.0183160

69 Bleuler, E. (1911). Dementia praecox oder de Gruppe der Schizophrenien. In G. Aschaffenburg (Ed.), Handbuch der Pscyhiatrie, 4 Abt. (pp. 251-252). Leipzig: Deuticke. Bolte, S., Duketis, E., Poustka, F., & Holtmann, M. (2011). Sex differences in cognitive domains and their clinical correlates in higher-functioning autism spectrum disorders. Autism, 15(4), 497-511. doi:10.1177/1362361310391116 Bolte, S., Poustka, F., & Constantino, J. N. (2008). Assessing autistic traits: cross-cultural validation of the social responsiveness scale (SRS). Autism Res, 1(6), 354-363. doi:10.1002/aur.49 Bosc, M. (2000). Assessment of social functioning in depression. Compr Psychiatry, 41(1), 63-69. doi:10.1016/s0010-440x(00)90133-0 Bourgeron, T. (2016). Current knowledge on the genetics of autism and propositions for future research. C R Biol, 339(7-8), 300-307. doi:10.1016/j.crvi.2016.05.004 Brugha, T. S., McManus, S., Bankart, J., Scott, F., Purdon, S., Smith, J., . . . Meltzer, H. (2011). Epidemiology of autism spectrum disorders in adults in the community in England. Arch Gen Psychiatry, 68(5), 459-465. doi:10.1001/archgenpsychiatry.2011.38 Buesher, A. V. S., Cidav, Z., Knapp, M., & Mandell, D. (2014). Costs of Autism Spectrum Disorders in the United Kingdom and the United States. JAMA Pediatrics, 168(8), 721-728. Cai, Q., Feng, L., & Yap, K. Z. (2018). Systematic review and meta-analysis of reported adverse events of long-term intranasal oxytocin treatment for autism spectrum disorder. Psychiatry Clin Neurosci, 72(3), 140-151. doi:10.1111/pcn.12627 Carlson, S. M., Claxton, L. J., & Moses, L. J. (2015). The Relation Between Executive Function and Theory of Mind is More Than Skin Deep. Journal of Cognition and Development, 16(1), 186-197. doi:10.1080/15248372.2013.824883 Chatham, C. H., Taylor, K. I., Charman, T., Liogier D'ardhuy, X., Eule, E., Fedele, A., . . . Bolognani, F. (2018). Adaptive behavior in autism: Minimal clinically important differences on the Vineland-II. Autism Res, 11(2), 270-283. doi:10.1002/aur.1874 Choi, S. W., Mak, T. S. H., & Reilly, P. F. (2018). A guide to performing Polygenic Risk Score analyses. bioRxiv preprint. doi:https://doi.org/10.1101/416545 Chouinard, B., Gallagher, L., & Kelly, C. (2019). He said, she said: Autism spectrum diagnosis and gender differentially affect relationships between executive functions and social communication. Autism, 1362361318815639. doi:10.1177/1362361318815639 Clarke, T. K., Lupton, M. K., Fernandez-Pujals, A. M., Starr, J., Davies, G., Cox, S., . . . McIntosh, A. M. (2016). Common polygenic risk for autism spectrum disorder (ASD) is associated with cognitive ability in the general population. Mol Psychiatry, 21(3), 419-425. doi:10.1038/mp.2015.12 Coghill, D. R., Seth, S., Pedroso, S., Usala, T., Currie, J., & Gagliano, A. (2014). Effects of methylphenidate on cognitive functions in children and adolescents with attention- deficit/hyperactivity disorder: evidence from a systematic review and a meta-analysis. Biol Psychiatry, 76(8), 603-615. doi:10.1016/j.biopsych.2013.10.005 Cohen, J. (1988). Statistical Power Analysis for the Behavioral Sciences (S. Edition Ed.). New York: Lawrence Erlbaum Associates. Cohen, J. (1992). A power primer. Psychol Bull, 112(1), 155-159. doi:10.1037//0033- 2909.112.1.155 Constantino, J. N. (2011). The quantitative nature of autistic social impairment. Pediatr Res, 69(5 Pt 2), 55R-62R. doi:10.1203/PDR.0b013e318212ec6e

70 Constantino, J. N., Davis, S. A., Todd, R. D., Schindler, M. K., Gross, M. M., Brophy, S. L., . . . Reich, W. (2003). Validation of a brief quantitative measure of autistic traits: comparison of the social responsiveness scale with the autism diagnostic interview- revised. J Autism Dev Disord, 33(4), 427-433. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/12959421 Constantino, J. N., & Gruber, C. P. (2005). Social Responsiveness Scale (SRS). Manual. United States of America: Western Psychological Service. Constantino, J. N., Gruber, C. P., Davis, S., Hayes, S., Passanante, N., & Przybeck, T. (2004). The factor structure of autistic traits. J Child Psychol Psychiatry, 45(4), 719-726. doi:10.1111/j.1469-7610.2004.00266.x Craig, F., Margari, F., Legrottaglie, A. R., Palumbi, R., de Giambattista, C., & Margari, L. (2016). A review of executive function deficits in autism spectrum disorder and attention-deficit/hyperactivity disorder. Neuropsychiatr Dis Treat, 12, 1191-1202. doi:10.2147/NDT.S104620 Das, S., Forer, L., Schonherr, S., Sidore, C., Locke, A. E., Kwong, A., . . . Fuchsberger, C. (2016). Next-generation genotype imputation service and methods. Nat Genet, 48(10), 1284-1287. doi:10.1038/ng.3656 Davidovitch, M., Hemo, B., Manning-Courtney, P., & Fombonne, E. (2013). Prevalence and incidence of autism spectrum disorder in an Israeli population. J Autism Dev Disord, 43(4), 785-793. doi:10.1007/s10803-012-1611-z de la Torre-Ubieta, L., Won, H., Stein, J. L., & Geschwind, D. H. (2016). Advancing the understanding of autism disease mechanisms through genetics. Nat Med, 22(4), 345- 361. doi:10.1038/nm.4071 De Luca, C. R., & Leventer, R. J. (2008). Developmental trajectories of executive functions across the lifespan. In V. Anderson, R. Jacobs, & P. J. Anderson (Eds.), Executive Functions and the Frontal Lobes: A Lifespan Perspective. New York: Taylor & Francis. de Vries, M., Prins, P. J., Schmand, B. A., & Geurts, H. M. (2015). Working memory and cognitive flexibility-training for children with an autism spectrum disorder: a randomized controlled trial. J Child Psychol Psychiatry, 56(5), 566-576. doi:10.1111/jcpp.12324 Dean, M., Harwood, R., & Kasari, C. (2017). The art of camouflage: Gender differences in the social behaviors of girls and boys with autism spectrum disorder. Autism, 21(6), 678-689. doi:10.1177/1362361316671845 Demetriou, E. A., Lampit, A., Quintana, D. S., Naismith, S. L., Song, Y. J. C., Pye, J. E., . . . Guastella, A. J. (2017). Autism spectrum disorders: a meta-analysis of executive function. Mol Psychiatry. doi:10.1038/mp.2017.75 Demontis, D., Walters, R. K., Martin, J., Mattheisen, M., Als, T. D., Agerbo, E., . . . Neale, B. M. (2018). Discovery of the first genome-wide significant risk loci for attention deficit/hyperactivity disorder. Nat Genet. doi:10.1038/s41588-018-0269-7 Dennett, D. C. (1978). Beliefs about beliefs. The Behavioral and Brain Sciences, 4, 568-570. Diamond, A. (2013). Executive functions. Annu Rev Psychol, 64, 135-168. doi:10.1146/annurev-psych-113011-143750 Doernberg, E., & Hollander, E. (2016). Neurodevelopmental Disorders (ASD and ADHD): DSM-5, ICD-10, and ICD-11. CNS Spectr, 21(4), 295-299. doi:10.1017/S1092852916000262 Dudbridge, F. (2013). Power and predictive accuracy of polygenic risk scores. PLoS Genet, 9(3), e1003348. doi:10.1371/journal.pgen.1003348

71 Dworzynski, K., Ronald, A., Bolton, P., & Happe, F. (2012). How different are girls and boys above and below the diagnostic threshold for autism spectrum disorders? J Am Acad Child Adolesc Psychiatry, 51(8), 788-797. doi:10.1016/j.jaac.2012.05.018 Editorial. (2010). A decade for psychiatric disorders. Nature, 463(7277), 9. doi:463009a [pii] 10.1038/463009a Elsabbagh, M., Divan, G., Koh, Y. J., Kim, Y. S., Kauchali, S., Marcin, C., . . . Fombonne, E. (2012). Global prevalence of autism and other pervasive developmental disorders. Autism Res, 5(3), 160-179. doi:10.1002/aur.239 Engel, G. L. (1977). The need for a new medical model: a challenge for biomedicine. Science, 196(4286), 129-136. doi:10.1126/science.847460 Evans, B. (2013). How autism became autism: The radical transformation of a central concept of child development in Britain. Hist Human Sci, 26(3), 3-31. doi:10.1177/0952695113484320 Fallmyr, Ø., & Egeland, J. (2011). Psykometriske egenskaper for den norske versjonen av Behavior Rating Inventory of Executive Function (BRIEF). Tidsskrift for Norsk Psykologforening, 48, 339-343. Fanous, A. H., Zhou, B., Aggen, S. H., Bergen, S. E., Amdur, R. L., Duan, J., . . . Schizophrenia Psychiatric Genome-Wide Association Study, C. (2012). Genome-wide association study of clinical dimensions of schizophrenia: polygenic effect on disorganized symptoms. Am J Psychiatry, 169(12), 1309-1317. doi:10.1176/appi.ajp.2012.12020218 Fombonne, E. (2018). Editorial: The rising prevalence of autism. J Child Psychol Psychiatry, 59(7), 717-720. doi:10.1111/jcpp.12941 Frazier, T. W., Georgiades, S., Bishop, S. L., & Hardan, A. Y. (2014). Behavioral and cognitive characteristics of females and males with autism in the Simons Simplex Collection. J Am Acad Child Adolesc Psychiatry, 53(3), 329-340 e321-323. doi:10.1016/j.jaac.2013.12.004 Friedman, N. P., Miyake, A., Corley, R. P., Young, S. E., Defries, J. C., & Hewitt, J. K. (2006). Not all executive functions are related to intelligence. Psychol Sci, 17(2), 172- 179. doi:10.1111/j.1467-9280.2006.01681.x Frith, U. (1989). Autism: Explaining the Enigma. Oxford, UK: Blackwell Publishers. Gardiner, E., & Iarocci, G. (2018). Everyday executive function predicts adaptive and internalizing behavior among children with and without autism spectrum disorder. Autism Res, 11(2), 284-295. doi:10.1002/aur.1877 Garon, N., Bryson, S. E., & Smith, I. M. (2008). Executive function in preschoolers: a review using an integrative framework. Psychol Bull, 134(1), 31-60. doi:10.1037/0033- 2909.134.1.31 Gaugler, T., Klei, L., Sanders, S. J., Bodea, C. A., Goldberg, A. P., Lee, A. B., . . . Buxbaum, J. D. (2014). Most genetic risk for autism resides with common variation. Nat Genet, 46(8), 881-885. doi:10.1038/ng.3039 Geschwind, D. H. (2011). Genetics of autism spectrum disorders. Trends Cogn Sci, 15(9), 409-416. doi:10.1016/j.tics.2011.07.003 Geurts, H. M., Corbett, B., & Solomon, M. (2009). The paradox of cognitive flexibility in autism. Trends Cogn Sci, 13(2), 74-82. doi:10.1016/j.tics.2008.11.006 Geurts, H. M., de Vries, M., & van den Bergh, S. F. (2014). Executive Functioning Theory and Autism. In S. Goldstein & J. A. Naglieri (Eds.), Handbook of Executive Functioning. New York: Springer Science + Business Media. Gillberg, C. (2010). The ESSENCE in child psychiatry: Early Symptomatic Syndromes Eliciting Neurodevelopmental Clinical Examinations. Res Dev Disabil, 31(6), 1543- 1551. doi:10.1016/j.ridd.2010.06.002

72 Gillberg, C., Allely, C., Bourgeron, T., Coleman, M., Fernell, E., Hadjikhani, N., & Sarovic, D. (2019). The Neurobiology of Autism. In F. R. Volkmar (Ed.), Autism and Pervasive Developmental Disorders. Cambridge: Cambridge University Press. Gilotty, L., Kenworthy, L., Sirian, L., Black, D. O., & Wagner, A. E. (2002). Adaptive skills and executive function in autism spectrum disorders. Child Neuropsychol, 8(4), 241- 248. doi:10.1076/chin.8.4.241.13504 Gioia, G. A., Isquith, P. K., Guy, S. C., & Kenworthy, L. (2000). Behavior Rating Rating Inventory of Executive Function (BRIEF). United States of America: PAR Psychological Assessment Resources, Inc. Gioia, G. A., Isquith, P. K., Guy, S. C., & Kenworthy, L. (2015). BRIEF 2. Behavior Rating Inventory of Executive Functions, Second Edition. Professional Manual. Florida: PAR. Gotham, K., Pickles, A., & Lord, C. (2009). Standardizing ADOS scores for a measure of severity in autism spectrum disorders. J Autism Dev Disord, 39(5), 693-705. doi:10.1007/s10803-008-0674-3 Grissom, N. M., & Reyes, T. M. (2019). Correction: Let's call the whole thing off: evaluating gender and sex differences in executive function. Neuropsychopharmacology, 44(7), 1344. doi:10.1038/s41386-019-0367-y Grove, J., Ripke, S., Als, T. D., Mattheisen, M., Walters, R. K., Won, H., . . . Borglum, A. D. (2019). Identification of common genetic risk variants for autism spectrum disorder. Nat Genet, 51(3), 431-444. doi:10.1038/s41588-019-0344-8 Halladay, A. K., Bishop, S., Constantino, J. N., Daniels, A. M., Koenig, K., Palmer, K., . . . Szatmari, P. (2015). Sex and gender differences in autism spectrum disorder: summarizing evidence gaps and identifying emerging areas of priority. Mol Autism, 6, 36. doi:10.1186/s13229-015-0019-y Happe, F. (1999). Autism: cognitive deficit or cognitive style? Trends Cogn Sci, 3(6), 216- 222. Retrieved from https://www.ncbi.nlm.nih.gov/pubmed/10354574 Havdahl, K. A., Bishop, S. L., Suren, P., Oyen, A. S., Lord, C., Pickles, A., . . . Stoltenberg, C. (2017). The influence of parental concern on the utility of autism diagnostic instruments. Autism Res, 10(10), 1672-1686. doi:10.1002/aur.1817 Helles, A., Gillberg, I. C., Gillberg, C., & Billstedt, E. (2017). Asperger syndrome in males over two decades: Quality of life in relation to diagnostic stability and psychiatric comorbidity. Autism, 21(4), 458-469. doi:10.1177/1362361316650090 Hill, E. L. (2004). Executive dysfunction in autism. Trends Cogn Sci, 8(1), 26-32. Hirvikoski, T., Mittendorfer-Rutz, E., Boman, M., Larsson, H., Lichtenstein, P., & Bolte, S. (2016). Premature mortality in autism spectrum disorder. Br J Psychiatry, 208(3), 232-238. doi:10.1192/bjp.bp.114.160192 Honey, E., Rodgers, J., & McConachie, H. (2012). Measurement of restricted and repetitive behaviour in children with autism spectrum disorder: Selecting a questionnaire or interview. Research in Autism Spectrum Disorders, 6(2), 757-776. Hovik, K. T., Egeland, J., Isquith, P. K., Gioia, G., Skogli, E. W., Andersen, P. N., & Oie, M. (2014). Distinct Patterns of Everyday Executive Function Problems Distinguish Children With Tourette Syndrome From Children With ADHD or Autism Spectrum Disorders. J Atten Disord. doi:10.1177/1087054714550336 Hoyland, A. L., Naerland, T., Engstrom, M., Lydersen, S., & Andreassen, O. A. (2017). The relation between face-emotion recognition and social function in adolescents with autism spectrum disorders: A case control study. PLoS One, 12(10), e0186124. doi:10.1371/journal.pone.0186124 Hutchison, S. M., Muller, U., & Iarocci, G. (2019). Parent Reports of Executive Function Associated with Functional Communication and Conversational Skills Among School

73 Age Children With and Without Autism Spectrum Disorder. J Autism Dev Disord. doi:10.1007/s10803-019-03958-6 International Schizophrenia Consortium, Purcell, S. M., Wray, N. R., Stone, J. L., Visscher, P. M., O'Donovan, M. C., . . . Sklar, P. (2009). Common polygenic variation contributes to risk of schizophrenia and bipolar disorder. Nature, 460(7256), 748-752. doi:10.1038/nature08185 Isquith, P. K., Roth, R. M., & Gioia, G. (2013). Contribution of rating scales to the assessment of executive functions. Appl Neuropsychol Child, 2(2), 125-132. doi:10.1080/21622965.2013.748389 Jackson, S. I. J., & Volkmar, F. R. (2019). Diagnosis and Definition of Autism and Other Pervasive Developmental Disorders. In F. R. Volkmar (Ed.), Autism and Pervasive Developmental Disorders (Third ed.). Cambridge: Cambridge University Press. Jacquemont, S., Coe, B. P., Hersch, M., Duyzend, M. H., Krumm, N., Bergmann, S., . . . Eichler, E. E. (2014). A higher mutational burden in females supports a "female protective model" in neurodevelopmental disorders. Am J Hum Genet, 94(3), 415-425. doi:10.1016/j.ajhg.2014.02.001 Jick, H., Beach, K. J., & Kaye, J. A. (2006). Incidence of autism over time. Epidemiology, 17(1), 120-121. doi:10.1097/01.ede.0000190553.72757.15 Johannessen, J., Naerland, T., Bloss, C., Rietschel, M., Strohmaier, J., Gjevik, E., . . . Andreassen, O. A. (2016). Parents' attitudes toward genetic research in autism spectrum disorder. Psychiatr Genet, 26(2), 74-80. doi:10.1097/YPG.0000000000000121 Johannessen, J., Naerland, T., Hope, S., Torske, T., Hoyland, A. L., Strohmaier, J., . . . Andreassen, O. A. (2017). Parents' Attitudes toward Clinical Genetic Testing for Autism Spectrum Disorder-Data from a Norwegian Sample. Int J Mol Sci, 18(5). doi:10.3390/ijms18051078 Joseph, R. M., McGrath, L. M., & Tager-Flusberg, H. (2005). Executive dysfunction and its relation to language ability in verbal school-age children with autism. Dev Neuropsychol, 27(3), 361-378. doi:10.1207/s15326942dn2703_4 Kanner, L. (1943). Autistic Disturbances of Affective Contact. Nervous Child(2), 217-250. Kanner, L. (1949). Problems of nosology and psychodynamics of early infantile autism. Am J Orthopsychiatry, 19(3), 416-426. doi:10.1111/j.1939-0025.1949.tb05441.x Kapp, S. K. (2018). Social Support, Well-being, and Quality of Life Among Individuals on the Autism Spectrum. Pediatrics, 141(Suppl 4), S362-S368. doi:10.1542/peds.2016- 4300N Kapp, S. K., Gillespie-Lynch, K., Sherman, L. E., & Hutman, T. (2013). Deficit, difference, or both? Autism and neurodiversity. Dev Psychol, 49(1), 59-71. doi:10.1037/a0028353 Karr, J. E., Areshenkoff, C. N., Rast, P., Hofer, S. M., Iverson, G. L., & Garcia-Barrera, M. A. (2018). The unity and diversity of executive functions: A systematic review and re- analysis of latent variable studies. Psychol Bull, 144(11), 1147-1185. doi:10.1037/bul0000160 Kenny, L., Cribb, S. J., & Pellicano, E. (2019). Childhood Executive Function Predicts Later Autistic Features and Adaptive Behavior in Young Autistic People: a 12-Year Prospective Study. J Abnorm Child Psychol, 47(6), 1089-1099. doi:10.1007/s10802- 018-0493-8 Kenworthy, L., Anthony, L. G., Naiman, D. Q., Cannon, L., Wills, M. C., Luong-Tran, C., . . . Wallace, G. L. (2014). Randomized controlled effectiveness trial of executive function intervention for children on the autism spectrum. J Child Psychol Psychiatry, 55(4), 374-383. doi:10.1111/jcpp.12161

74 Kenworthy, L., Black, D. O., Harrison, B., della Rosa, A., & Wallace, G. L. (2009). Are executive control functions related to autism symptoms in high-functioning children? Child Neuropsychol, 15(5), 425-440. doi:10.1080/09297040802646983 Kenworthy, L., Yerys, B. E., Anthony, L. G., & Wallace, G. L. (2008). Understanding executive control in autism spectrum disorders in the lab and in the real world. Neuropsychol Rev, 18(4), 320-338. doi:10.1007/s11065-008-9077-7 Kim, J. Y., Son, M. J., Son, C. Y., Radua, J., Eisenhut, M., Gressier, F., . . . Fusar-Poli, P. (2019). Environmental risk factors and biomarkers for autism spectrum disorder: an umbrella review of the evidence. Lancet Psychiatry, 6(7), 590-600. doi:10.1016/S2215-0366(19)30181-6 Kleinhans, N., Akshoomoff, N., & Delis, D. C. (2005). Executive functions in autism and Asperger's disorder: flexibility, fluency, and inhibition. Dev Neuropsychol, 27(3), 379- 401. doi:10.1207/s15326942dn2703_5 Kouklari, E. C., Tsermentseli, S., & Monks, C. P. (2018). Hot and cool executive function in children and adolescents with autism spectrum disorder: Cross-sectional developmental trajectories. Child Neuropsychol, 24(8), 1088-1114. doi:10.1080/09297049.2017.1391190 Kouklari, E. C., Tsermentseli, S., & Monks, C. P. (2019). Developmental trends of hot and cool executive function in school-aged children with and without autism spectrum disorder: Links with theory of mind. Dev Psychopathol, 31(2), 541-556. doi:10.1017/S0954579418000081 Lai, M. C., Lombardo, M. V., Auyeung, B., Chakrabarti, B., & Baron-Cohen, S. (2015). Sex/gender differences and autism: setting the scene for future research. J Am Acad Child Adolesc Psychiatry, 54(1), 11-24. doi:10.1016/j.jaac.2014.10.003 Lai, M. C., Lombardo, M. V., & Baron-Cohen, S. (2014). Autism. Lancet, 383(9920), 896- 910. doi:10.1016/s0140-6736(13)61539-1 Lai, M. C., Lombardo, M. V., Ruigrok, A. N., Chakrabarti, B., Auyeung, B., Szatmari, P., . . . Consortium, M. A. (2017). Quantifying and exploring camouflaging in men and women with autism. Autism, 21(6), 690-702. doi:10.1177/1362361316671012 Lai, M. C., Lombardo, M. V., Suckling, J., Ruigrok, A. N., Chakrabarti, B., Ecker, C., . . . Baron-Cohen, S. (2013). Biological sex affects the neurobiology of autism. Brain, 136(Pt 9), 2799-2815. doi:10.1093/brain/awt216 Landa, R. J., & Goldberg, M. C. (2005). Language, social, and executive functions in high functioning autism: a continuum of performance. J Autism Dev Disord, 35(5), 557- 573. doi:10.1007/s10803-005-0001-1 Lehnhardt, F. G., Falter, C. M., Gawronski, A., Pfeiffer, K., Tepest, R., Franklin, J., & Vogeley, K. (2016). Sex-Related Cognitive Profile in Autism Spectrum Disorders Diagnosed Late in Life: Implications for the Female Autistic Phenotype. J Autism Dev Disord, 46(1), 139-154. doi:10.1007/s10803-015-2558-7 Lemon, J. M., Gargaro, B., Enticott, P. G., & Rinehart, N. J. (2011). Executive functioning in autism spectrum disorders: a gender comparison of response inhibition. J Autism Dev Disord, 41(3), 352-356. doi:10.1007/s10803-010-1039-2 Leung, R. C., Vogan, V. M., Powell, T. L., Anagnostou, E., & Taylor, M. J. (2015). The role of executive functions in social impairment in Autism Spectrum Disorder. Child Neuropsychol, 1-9. doi:10.1080/09297049.2015.1005066 Levy, D., Ronemus, M., Yamrom, B., Lee, Y. H., Leotta, A., Kendall, J., . . . Wigler, M. (2011). Rare de novo and transmitted copy-number variation in autistic spectrum disorders. Neuron, 70(5), 886-897. doi:10.1016/j.neuron.2011.05.015 Lewis, C. M., & Vassos, E. (2017). Prospects for using risk scores in polygenic medicine. Genome Med, 9(1), 96. doi:10.1186/s13073-017-0489-y

75 Lezak, M. D., Howieson, D. B., & Loring, D. W. (2004). Basic Concepts. In M. D. Lezak, D. B. Howieson, & D. W. Loring (Eds.), Neurpsychological Assessment (Fourth Edition ed.). New York: Oxford University Press. Li, Y., Willer, C. J., Ding, J., Scheet, P., & Abecasis, G. R. (2010). MaCH: using sequence and genotype data to estimate haplotypes and unobserved genotypes. Genet Epidemiol, 34(8), 816-834. doi:10.1002/gepi.20533 Loe, I. M., Chatav, M., & Alduncin, N. (2015). Complementary assessments of executive function in preterm and full-term preschoolers. Child Neuropsychol, 21(3), 331-353. doi:10.1080/09297049.2014.906568 Loomes, R., Hull, L., & Mandy, W. P. L. (2017). What Is the Male-to-Female Ratio in Autism Spectrum Disorder? A Systematic Review and Meta-Analysis. J Am Acad Child Adolesc Psychiatry, 56(6), 466-474. doi:10.1016/j.jaac.2017.03.013 Lord, C. (2019). Recognising the heterogeneity of autism. Lancet Psychiatry, 6(7), 551-552. doi:10.1016/S2215-0366(19)30220-2 Lord, C., Elsabbagh, M., Baird, G., & Veenstra-Vanderweele, J. (2018). Autism spectrum disorder. Lancet, 392(10146), 508-520. doi:10.1016/S0140-6736(18)31129-2 Lord, C., Risi, S., Lambrecht, L., Cook, E. H., Jr., Leventhal, B. L., DiLavore, P. C., . . . Rutter, M. (2000). The autism diagnostic observation schedule-generic: a standard measure of social and communication deficits associated with the spectrum of autism. J Autism Dev Disord, 30(3), 205-223. Retrieved from https://www.ncbi.nlm.nih.gov/pubmed/11055457 Lord, C., Rutter, M., & Le Couteur, A. (1994). Autism Diagnostic Interview-Revised: a revised version of a diagnostic interview for caregivers of individuals with possible pervasive developmental disorders. J Autism Dev Disord, 24(5), 659-685. Retrieved from https://www.ncbi.nlm.nih.gov/pubmed/7814313 Lyall, K., Croen, L., Daniels, J., Fallin, M. D., Ladd-Acosta, C., Lee, B. K., . . . Newschaffer, C. (2017). The Changing Epidemiology of Autism Spectrum Disorders. Annu Rev Public Health, 38, 81-102. doi:10.1146/annurev-publhealth-031816-044318 Magiati, I., & Howlin, P. (2019). Adult Life for People with Autism Spectrum Disorders. In F. R. Volkmar (Ed.), Autism and Pervasive Developmental Disorders (Third Edition ed.). Cambridge: Cambridge University Press. Mahone, E. M., Cirino, P. T., Cutting, L. E., Cerrone, P. M., Hagelthorn, K. M., Hiemenz, J. R., . . . Denckla, M. B. (2002). Validity of the behavior rating inventory of executive function in children with ADHD and/or Tourette syndrome. Arch Clin Neuropsychol, 17(7), 643-662. Manolio, T. A., Collins, F. S., Cox, N. J., Goldstein, D. B., Hindorff, L. A., Hunter, D. J., . . . Visscher, P. M. (2009). Finding the missing heritability of complex diseases. Nature, 461(7265), 747-753. doi:10.1038/nature08494 Martin, A. R., Daly, M. J., Robinson, E. B., Hyman, S. E., & Neale, B. M. (2018). Predicting Polygenic Risk of Psychiatric Disorders. Biol Psychiatry. doi:10.1016/j.biopsych.2018.12.015 Martin, A. R., Kanai, M., Kamatani, Y., Okada, Y., Neale, B. M., & Daly, M. J. (2019). Clinical use of current polygenic risk scores may exacerbate health disparities. Nat Genet, 51(4), 584-591. doi:10.1038/s41588-019-0379-x Masi, A., DeMayo, M. M., Glozier, N., & Guastella, A. J. (2017). An Overview of Autism Spectrum Disorder, Heterogeneity and Treatment Options. Neurosci Bull, 33(2), 183- 193. doi:10.1007/s12264-017-0100-y Matsuura, N., Ishitobi, M., Arai, S., Kawamura, K., Asano, M., Inohara, K., . . . Kosaka, H. (2014). Distinguishing between autism spectrum disorder and attention deficit hyperactivity disorder by using behavioral checklists, cognitive assessments, and

76 neuropsychological test battery. Asian J Psychiatr, 12, 50-57. doi:10.1016/j.ajp.2014.06.011 McAuley, T., Chen, S., Goos, L., Schachar, R., & Crosbie, J. (2010). Is the behavior rating inventory of executive function more strongly associated with measures of impairment or executive function? J Int Neuropsychol Soc, 16(3), 495-505. doi:10.1017/S1355617710000093 Merchan-Naranjo, J., Boada, L., del Rey-Mejias, A., Mayoral, M., Llorente, C., Arango, C., & Parellada, M. (2016). Executive function is affected in autism spectrum disorder, but does not correlate with intelligence. Rev Psiquiatr Salud Ment, 9(1), 39-50. doi:10.1016/j.rpsm.2015.10.005 Miyake, A., Friedman, N. P., Emerson, M. J., Witzki, A. H., Howerter, A., & Wager, T. D. (2000). The unity and diversity of executive functions and their contributions to complex "Frontal Lobe" tasks: a latent variable analysis. Cogn Psychol, 41(1), 49-100. doi:10.1006/cogp.1999.0734 Mooney, L., Fosdick, C., & Erickson, C. A. (2019). Psychopharmacology of Autism Spectrum Disorders. In F. R. Volkmar (Ed.), Autism and Pervasive Developmental Disorders (Third ed.). Cambridge: Cambridge University Press. Myers, J., Chavez, A., Hill, A. P., Zuckerman, K., & Fombonne, E. (2019). Epidemiological Surveys of Autism Spectrum Disorders. In F. R. Volkmar (Ed.), Autism and Pervasive Developmental Disorders: Cambridge University PressS. Nahmias, A. S., Pellecchia, M., Stahmer, A. C., & Mandell, D. S. (2019). Effectiveness of community-based early intervention for children with autism spectrum disorder: a meta-analysis. J Child Psychol Psychiatry. doi:10.1111/jcpp.13073 National Insitute of Health. (2020). What is precision medicine? 11th of February. Retrieved from https://ghr.nlm.nih.gov/primer/precisionmedicine/definition Nelson, A. T., Lopata, C., Volker, M. A., Thomeer, M. L., Toomey, J. A., & Dua, E. (2016). Exploratory Factor Analysis of SRS-2 Teacher Ratings for Youth with ASD. J Autism Dev Disord, 46(9), 2905-2915. doi:10.1007/s10803-016-2822-5 Odom, S. L., Morin, K., Savage, M., & Tomaszewski, B. (2019). Behavioral and Educationsal Interventions. In F. R. Volkmar (Ed.), Autism and Pervasive Developmental Disorders. Cambridge: Cambridge University Press. Oxford Dictionary. (2020). Cognition - definition of cognition from the Oxford Dictionary. 9th of February. Retrieved from https://www.lexico.com/definition/cognition Pagnozzi, A. M., Conti, E., Calderoni, S., Fripp, J., & Rose, S. E. (2018). A systematic review of structural MRI biomarkers in autism spectrum disorder: A machine learning perspective. Int J Dev Neurosci, 71, 68-82. doi:10.1016/j.ijdevneu.2018.08.010 Palk, A. C., Dalvie, S., de Vries, J., Martin, A. R., & Stein, D. J. (2019). Potential use of clinical polygenic risk scores in psychiatry - ethical implications and communicating high polygenic risk. Philos Ethics Humanit Med, 14(1), 4. doi:10.1186/s13010-019- 0073-8 Parellada, M., Penzol, M. J., Pina, L., Moreno, C., Gonzalez-Vioque, E., Zalsman, G., & Arango, C. (2014). The neurobiology of autism spectrum disorders. Eur Psychiatry, 29(1), 11-19. doi:10.1016/j.eurpsy.2013.02.005 Paul, R., & Simmons, E. S. (2019). Communication and its Development in Autism Spectrum Disorders. In F. R. Volkmar (Ed.), Autism and Pervasive Developmental Disorders. Cambridge: Cambridge University Press. Pellicano, E. (2010). Individual differences in executive function and central coherence predict developmental changes in theory of mind in autism. Dev Psychol, 46(2), 530- 544. doi:10.1037/a0018287

77 Pellicano, E. (2011). Psychological models of autism: an overview. In I. Roth & P. Rezaie (Eds.), Researching the Autism Spectrum: Contemporary Perspectives Cambridge Univeristy Press. Pennington, B. F., & Ozonoff, S. (1996). Executive functions and developmental psychopathology. J Child Psychol Psychiatry, 37(1), 51-87. Puhr, A., Ruud, E., Anderson, V., Due-Tonnesen, B. J., Skarbo, A. B., Finset, A., & Andersson, S. (2019). Self-Reported Executive Dysfunction, Fatigue, and Psychological and Emotional Symptoms in Physically Well-Functioning Long-Term Survivors of Pediatric Brain Tumor. Dev Neuropsychol, 44(1), 88-103. doi:10.1080/87565641.2018.1540007 Purcell, S., Neale, B., Todd-Brown, K., Thomas, L., Ferreira, M. A., Bender, D., . . . Sham, P. C. (2007). PLINK: a tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet, 81(3), 559-575. doi:10.1086/519795 Quintana, D. S., Westlye, L. T., Hope, S., Naerland, T., Elvsashagen, T., Dorum, E., . . . Andreassen, O. A. (2017). Dose-dependent social-cognitive effects of intranasal oxytocin delivered with novel Breath Powered device in adults with autism spectrum disorder: a randomized placebo-controlled double-blind crossover trial. Transl Psychiatry, 7(5), e1136. doi:10.1038/tp.2017.103 Ramaswami, G., & Geschwind, D. H. (2018). Genetics of autism spectrum disorder. Handb Clin Neurol, 147, 321-329. doi:10.1016/B978-0-444-63233-3.00021-X Rao, T. S., & Andrade, C. (2011). The MMR vaccine and autism: Sensation, refutation, retraction, and fraud. Indian J Psychiatry, 53(2), 95-96. doi:10.4103/0019-5545.82529 Reiersen, A. M., & Todd, R. D. (2008). Co-occurrence of ADHD and autism spectrum disorders: phenomenology and treatment. Expert Rev Neurother, 8(4), 657-669. doi:10.1586/14737175.8.4.657 Robinson, K. E., Kaizar, E., Catroppa, C., Godfrey, C., & Yeates, K. O. (2014). Systematic review and meta-analysis of cognitive interventions for children with central nervous system disorders and neurodevelopmental disorders. J Pediatr Psychol, 39(8), 846- 865. doi:10.1093/jpepsy/jsu031 Roca, M., Manes, F., Cetkovich, M., Bruno, D., Ibanez, A., Torralva, T., & Duncan, J. (2014). The relationship between executive functions and fluid intelligence in schizophrenia. Front Behav Neurosci, 8, 46. doi:10.3389/fnbeh.2014.00046 Rommelse, N. N., Geurts, H. M., Franke, B., Buitelaar, J. K., & Hartman, C. A. (2011). A review on cognitive and brain endophenotypes that may be common in autism spectrum disorder and attention-deficit/hyperactivity disorder and facilitate the search for pleiotropic genes. Neurosci Biobehav Rev, 35(6), 1363-1396. doi:10.1016/j.neubiorev.2011.02.015 Rutter, M., Bailey, A., & Lord, C. (2003). The Social Communication Questionnaire (SCQ). Los Angeles: Western Psychological Services. Rutter, M., Lord, C., & LeCouteur, A. (2016). Autism Diagnostic Interview – Revised. Norsk versjon. Stockholm: Hogrefe Psykologforlag AB. Savage, J. E., Jansen, P. R., Stringer, S., Watanabe, K., Bryois, J., de Leeuw, C. A., . . . Posthuma, D. (2018). Genome-wide association meta-analysis in 269,867 individuals identifies new genetic and functional links to intelligence. Nat Genet, 50(7), 912-919. doi:10.1038/s41588-018-0152-6 Schork, A. J., Brown, T. T., Hagler, D. J., Thompson, W. K., Chen, C. H., Dale, A. M., . . . Genetics, S. (2018). Polygenic risk for psychiatric disorders correlates with executive function in typical development. Genes Brain Behav, e12480. doi:10.1111/gbb.12480 Simonoff, E., Pickles, A., Charman, T., Chandler, S., Loucas, T., & Baird, G. (2008). Psychiatric disorders in children with autism spectrum disorders: prevalence,

78 comorbidity, and associated factors in a population-derived sample. J Am Acad Child Adolesc Psychiatry, 47(8), 921-929. doi:10.1097/CHI.0b013e318179964f Skogli, E. W., Teicher, M. H., Andersen, P. N., Hovik, K. T., & Oie, M. (2013). ADHD in girls and boys--gender differences in co-existing symptoms and executive function measures. BMC Psychiatry, 13, 298. doi:10.1186/1471-244X-13-298 Smolker, H. R., Friedman, N. P., Hewitt, J. K., & Banich, M. T. (2018). Neuroanatomical Correlates of the Unity and Diversity Model of Executive Function in Young Adults. Front Hum Neurosci, 12, 283. doi:10.3389/fnhum.2018.00283 Snyder, H. R., Miyake, A., & Hankin, B. L. (2015). Advancing understanding of executive function impairments and psychopathology: bridging the gap between clinical and cognitive approaches. Front Psychol, 6, 328. doi:10.3389/fpsyg.2015.00328 Sonderby, I. E., Gustafsson, O., Doan, N. T., Hibar, D. P., Martin-Brevet, S., Abdellaoui, A., . . . p11.2 European Consortium, f. t. E.-C. N. V. w. g. (2018). Dose response of the 16p11.2 distal copy number variant on intracranial volume and basal ganglia. Mol Psychiatry. doi:10.1038/s41380-018-0118-1 Stanfield, A. C., McIntosh, A. M., Spencer, M. D., Philip, R., Gaur, S., & Lawrie, S. M. (2008). Towards a neuroanatomy of autism: a systematic review and meta-analysis of structural magnetic resonance imaging studies. Eur Psychiatry, 23(4), 289-299. doi:10.1016/j.eurpsy.2007.05.006 Sullivan, P. F., & Geschwind, D. H. (2019). Defining the Genetic, Genomic, Cellular, and Diagnostic Architectures of Psychiatric Disorders. Cell, 177(1), 162-183. doi:10.1016/j.cell.2019.01.015 Sun, X., Allison, C., Wei, L., Matthews, F. E., Auyeung, B., Wu, Y. Y., . . . Brayne, C. (2019). Autism prevalence in China is comparable to Western prevalence. Mol Autism, 10, 7. doi:10.1186/s13229-018-0246-0 Suren, P., Havdahl, A., Oyen, A. S., Schjolberg, S., Reichborn-Kjennerud, T., Magnus, P., . . . Stoltenberg, C. (2019). Diagnosing autism spectrum disorder among children in Norway. Tidsskr Nor Laegeforen. doi:10.4045/tidsskr.18.0960 Tamm, L., & Nakonezny, P. A. (2015). Metacognitive executive function training for young children with ADHD: a proof-of-concept study. Atten Defic Hyperact Disord, 7(3), 183-190. doi:10.1007/s12402-014-0162-x Thapar, A., & Cooper, M. (2016). Attention deficit hyperactivity disorder. Lancet, 387(10024), 1240-1250. doi:10.1016/S0140-6736(15)00238-X Tidmarsh, L., & Volkmar, F. R. (2003). Diagnosis and epidemiology of autism spectrum disorders. Can J Psychiatry, 48(8), 517-525. doi:10.1177/070674370304800803 Tiego, J., Bellgrove, M. A., Whittle, S., Pantelis, C., & Testa, R. (2019). Common mechanisms of executive attention underlie executive function and effortful control in children. Dev Sci, e12918. doi:10.1111/desc.12918 Tillmann, J., San Jose Caceres, A., Chatham, C. H., Crawley, D., Holt, R., Oakley, B., . . . group, E.-A. L. (2019). Investigating the factors underlying adaptive functioning in autism in the EU-AIMS Longitudinal European Autism Project. Autism Res, 12(4), 645-657. doi:10.1002/aur.2081 Toplak, M. E., West, R. F., & Stanovich, K. E. (2013). Practitioner review: do performance- based measures and ratings of executive function assess the same construct? J Child Psychol Psychiatry, 54(2), 131-143. doi:10.1111/jcpp.12001 Torkamani, A., Wineinger, N. E., & Topol, E. J. (2018). The personal and clinical utility of polygenic risk scores. Nat Rev Genet, 19(9), 581-590. doi:10.1038/s41576-018-0018-x Uddin, M., Wang, Y., & Woodbury-Smith, M. (2019). Artificial intelligence for precision medicine in neurodevelopmental disorders. NPJ Digit Med, 2, 112. doi:10.1038/s41746-019-0191-0

79 Van Eylen, L., Boets, B., Steyaert, J., Wagemans, J., & Noens, I. (2015). Executive functioning in autism spectrum disorders: influence of task and sample characteristics and relation to symptom severity. Eur Child Adolesc Psychiatry, 24(11), 1399-1417. doi:10.1007/s00787-015-0689-1 Vriezen, E. R., & Pigott, S. E. (2002). The relationship between parental report on the BRIEF and performance-based measures of executive function in children with moderate to severe traumatic brain injury. Child Neuropsychol, 8(4), 296-303. doi:10.1076/chin.8.4.296.13505 Wang, M., Li, K., Zhao, D., & Li, L. (2017). The association between maternal use of folic acid supplements during pregnancy and risk of autism spectrum disorders in children: a meta-analysis. Mol Autism, 8, 51. doi:10.1186/s13229-017-0170-8 Wechsler, D. (2002). Wechsler Preschool and Primary Scale of Intelligence-Third edition. . San Antonio, TX: Pearson. Wechsler, D. (2003). Wechler Intelligence Scale for Children- Fourth edition. San Antonio, TX: Pearson. Wechsler, D. (2008). Wechsler Adult Intelligence Scale–Fourth Edition. San Antonio, TX: Pearson. Wells, A., & Cartwright-Hatton, S. (2004). A short form of the metacognitions questionnaire: properties of the MCQ-30. Behav Res Ther, 42(4), 385-396. doi:10.1016/S0005- 7967(03)00147-5 Werling, D. M. (2016). The role of sex-differential biology in risk for autism spectrum disorder. Biol Sex Differ, 7, 58. doi:10.1186/s13293-016-0112-8 Werling, D. M., & Geschwind, D. H. (2013). Understanding sex bias in autism spectrum disorder. Proc Natl Acad Sci U S A, 110(13), 4868-4869. doi:10.1073/pnas.1301602110 Werling, D. M., Parikshak, N. N., & Geschwind, D. H. (2016). Gene expression in human brain implicates sexually dimorphic pathways in autism spectrum disorders. Nat Commun, 7, 10717. doi:10.1038/ncomms10717 White, E. I., Wallace, G. L., Bascom, J., Armour, A. C., Register-Brown, K., Popal, H. S., . . . Kenworthy, L. (2017). Sex differences in parent-reported executive functioning and adaptive behavior in children and young adults with autism spectrum disorder. Autism Res, 10(10), 1653-1662. doi:10.1002/aur.1811 White, S. J. (2013). The Triple I Hypothesis: taking another('s) perspective on executive dysfunction in autism. J Autism Dev Disord, 43(1), 114-121. doi:10.1007/s10803-012- 1550-8 Wigham, S., McConachie, H., Tandos, J., Le Couteur, A. S., & Gateshead Millennium Study core, t. (2012). The reliability and validity of the Social Responsiveness Scale in a UK general child population. Res Dev Disabil, 33(3), 944-950. doi:10.1016/j.ridd.2011.12.017 Wing, L. (1981). Asperger's syndrome: a clinical account. Psychol Med, 11(1), 115-129. doi:10.1017/s0033291700053332 World Health Organization. (1992). The ICD-10 classification of mental and behavioral disorders: clinical descriptions and diagnostic guidelines. Geneva: World Health Organization. Wray, N. R., Ripke, S., Mattheisen, M., Trzaskowski, M., Byrne, E. M., Abdellaoui, A., . . . Major Depressive Disorder Working Group of the Psychiatric Genomics, C. (2018). Genome-wide association analyses identify 44 risk variants and refine the genetic architecture of major depression. Nat Genet, 50(5), 668-681. doi:10.1038/s41588-018- 0090-3

80 Zelazo, P. D., & Carlson, S. M. (2012). Hot and cool executive function in childhood and adolescence: Developmental and plasticity. Child Developmen Perspectives(6), 354- 360. Zheutling, A. B., & Ross, D. (2018). Polygenic Risk Scores: What Are They Good For? Biological Psychiatry, 83(11), e51-e53. doi:10.1016/j.biopsych.2018.04.007 Zimmerman, D. L., Ownsworth, T., O'Donovan, A., Roberts, J., & Gullo, M. J. (2016). Independence of Hot and Cold Executive Function Deficits in High-Functioning Adults with Autism Spectrum Disorder. Front Hum Neurosci, 10, 24. doi:10.3389/fnhum.2016.00024 Ørbeck, B. (2009). Sosial Responsivitets Skala. Norsk oversettelse.: Hogrefe Psykologiske Forlag.

81

I

ORIGINAL RESEARCH published: 10 January 2018 doi: 10.3389/fnbeh.2017.00258

Metacognitive Aspects of Executive Function Are Highly Associated with Social Functioning on Parent-Rated Measures in Children with Autism Spectrum Disorder

Tonje Torske 1*, Terje Nærland 2, 3 , Merete G. Øie 4, 5 , Nina Stenberg 6 and Ole A. Andreassen 3

1 Division of Mental Health and Addiction, Vestre Viken Hospital Trust, Drammen, Norway, 2 NevSom Department of Rare Disorders and Disabilities, Oslo University Hospital, Oslo, Norway, 3 NORMENT, KG Jebsen Centre for Psychosis Research, University of Oslo and Oslo University Hospital, Oslo, Norway, 4 Department of Psychology, University of Oslo, Oslo, Norway, 5 Research Department, Innlandet Hospital Trust, Lillehammer, Norway, 6 Division of Mental Health and Addiction, Oslo University Hospital, Oslo, Norway

Autism Spectrum Disorder (ASD) is characterized by social dysfunction. Even though executive dysfunction has been recognized as important in understanding ASD, the findings are inconsistent. This might be due to different definitions of executive function (EF), which part of EF that has been studied, structured vs. unstructured tasks, inclusion Edited by: Lynne Ann Barker, of different moderators (age, IQ, sex) and different diagnostic categories within the Sheffield Hallam University, spectrum. The main finding is that people with ASD have more EF difficulties than United Kingdom normal controls and more difficulties on open-end tasks than on structured cognitive Reviewed by: Christos Frantzidis, tasks. Since some EF difficulties may not be observable in a laboratory setting, informant Aristotle University of Thessaloniki, measures might have higher ecological validity than neuropsychological tests. Evidence Greece suggests that executive dysfunctions are associated with social impairments, but few Margot J. Taylor, Hospital for Sick Children, Canada studies have investigated the details of this relationship, and it remains unclear what

*Correspondence: types of EF deficits are relevant for the social problems of individuals with ASD. Here we Tonje Torske investigated which EF domains were associated with various domains of social function [email protected] on parent-rated measures. A total of 86 children and adolescents with a diagnosis of ASD

Received: 15 March 2017 were included and tested for general cognitive abilities. Parents completed the Behavior Accepted: 19 December 2017 Rating Inventory of Executive Function (BRIEF) and the Social Responsiveness Scale Published: 10 January 2018 (SRS). Multiple regression analysis revealed significant associations between SRS scores Citation: Torske T, Nærland T, Øie MG, and age, sex, total IQ and the BRIEF indexes. The Metacognition Index from the BRIEF Stenberg N and Andreassen OA added significantly to the prediction of the SRS total score and the subscales Social (2018) Metacognitive Aspects of Communication, Social Motivation and Autistic Mannerisms. The findings suggest that Executive Function Are Highly Associated with Social Functioning on metacognitive aspects of EF are of particular importance for social abilities in children and Parent-Rated Measures in Children adolescents with ASD. Earlier research has shown that typically developing (TD) children with Autism Spectrum Disorder. Front. Behav. Neurosci. 11:258. have a different relationship between EF and social function than children with ASD. doi: 10.3389/fnbeh.2017.00258 They found that in TD children the EF domain related to behavioral regulation was most

Frontiers in Behavioral Neuroscience | www.frontiersin.org 1 January 2018 | Volume 11 | Article 258 Torske et al. Executive and Social Function in ASD

important to social function. The results from the current study may have implications for understanding the cognitive components of the social problems that define ASD, and may be relevant in developing more targeted clinical EF interventions related to core ASD dysfunctions.

Keywords: executive function, social function, autism spectrum disorder (ASD), behavior rating inventory of executive function, social responsiveness scale

INTRODUCTION (Gioia et al., 2000). A frequently used scale is the Behavior Rating Inventory of Executive Function (BRIEF) (Gioia et al., Autism Spectrum Disorder (ASD) is characterized with persistent 2000). The BRIEF is divided into a Behavioral Regulation Index deficits in social communication and social interaction across (BRI), and a Metacognition Index (MI), which together form multiple contexts, and restricted, repetitive patterns of behavior, aGlobalExecutiveComposite(GEC).TheBRIcomprisesthe interests, or activities (American Psychiatric Association, 2013). child’s ability to modulate both behavior and emotional control, Executive function (EF) deficits are common in children with and the ability to move flexible from one activity to another. ASD (Pennington and Ozonoff, 1996; Hill, 2004; Geurts et al., The MI is related to the child’s ability for active problem solving, 2014a), but not part of the diagnostic criteria. Furthermore, and to initiate, organize and monitor their own actions (Gioia EF correlates strongly with adaptive behavior (Gilotty et al., et al., 2000). Deficits in metacognitive aspects of EF (MI) have in 2002)andinfluencesQualityofLife(QoL)inchildrenwith earlier studies shown to be of particular importance to adaptive ASD (de Vries and Geurts, 2015). EF is often defined as functioning in high functioning children with ASD (Gilotty et al., the process of physical, cognitive, and emotional self-control 2002). and self-regulation that are necessary to maintain an effective While both social and EF deficits in ASD have been extensively goal-directed behavior (Pennington and Ozonoff, 1996; Corbett studied separately, there has been limited research on the et al., 2009; Diamond, 2013). EF comprises several components relationship between EF and social function. Some findings including inhibition, working memory, flexibility, emotional underpin the relationship between EF and key social concepts control, initiation, planning, organization, and self-control in ASD like joint attention and Theory of Mind (ToM). Dawson (Miyake et al., 2000; Hill, 2004). Even though executive (Dawson et al., 2002) argues that performance on ventromedial dysfunction has been recognized as important in understanding prefrontal EF tasks is strongly correlated to joint attention ability ASD, the findings are inconsistent (Van Eylen et al., 2015). One in young children. Joint attention is an important prerequisite explanation might be that EF is an umbrella term comprising for social functioning and often impaired in ASD (Dawson several components, and researchers have focused on different et al., 2002). Pellicano found that individual differences in subdomains. Meta-analyses and reviews have been written about EF in early life, predicted change in children’s ToM skills domains like inhibition and interference control (Geurts et al., (Pellicano, 2010). This line of research has to a large extent 2014b), cognitive flexibility (Leung and Zakzanis, 2014)and been based on laboratory measures and neuropsychological working memory in ASD (Barendse et al., 2013; Wang et al., test results. Leung et al. explored the role of EF in social 2017). All conclude that patients with ASD have EF deficits impairment in ASD using informant-based measures (Leung within the investigated areas, but not all found differences et al., 2015). They reported that both BRI and MI from the between ASD and normal controls on neuropsychological BRIEF predicted social functioning in children with ASD, while testing. Furthermore, the most consistent and striking difficulties only BRI predicted social functioning in the general population. are seen on tasks that are open-ended in structure, lack explicit In contrast, Kenworthy et al. (2009) found that EF, such as instructions and involve arbitrary rules (White, 2013; Van behavior regulation from BRIEF and semantic fluency and Eylen et al., 2015). Therefore, some of the inconsistency is divided auditory attention, correlated with autistic symptoms. suggested to be due to different types of measurements (parent- In an objective neuropsychological assessment of EF in ASD rated measures vs. neuropsychological testing). Individuals with children, Landa and Goldberg (2005) found no relationship ASD often display pronounced EF deficits in daily life, while between EF and social skills. performing adequately on highly structured neuropsychological Due to the heterogeneity of ASD (Lai et al., 2014), tasks (Kenworthy et al., 2008). The presence of comorbidities like more insight may also be gained from studying the range Attention Deficit Hyperactivity Disorder (ADHD) also increase of social difficulties beyond diagnostic categories. The Social the risk of EF difficulties (Craig et al., 2016). Most research Responsiveness Scale (SRS) (Constantino and Gruber, 2005), has focused on data from neuropsychological assessments of EF is a questionnaire designed to identify the presence and and/or how EF impairment is related to a diagnosis of ASD (Hill, severity of social impairments associated with ASD. The SRS 2004; Leung et al., 2015). Since some EF difficulties may not consists of five subscales which were developed to differentiate be observable in a laboratory setting, informant measures might between the subcategories of social impairments for children have higher ecological validity than neuropsychological tests with ASD (Constantino and Gruber, 2005). This continuous (Kenworthy et al., 2008). For this reason, questionnaires have scale allows trait quantification and focuses on functions closer been developed to investigate EF deficits in everyday life settings linked to activities in everyday life that diagnostic tools might

Frontiers in Behavioral Neuroscience | www.frontiersin.org 2 January 2018 | Volume 11 | Article 258 Torske et al. Executive and Social Function in ASD fail to capture (Achenbach, 2011; Nelson et al., 2016). Most The aim of the present study was to explore the relationship findings support a one-factor model of SRS (Constantino et al., between social functioning measured with the SRS (total score 2004; Bolte et al., 2008), while others have found evidence and the subscales Social Awareness, Social Cognition, Social for several dimensions (Nelson et al., 2016). Although most Communication, Social Motivation and Autistic Mannerisms) research with the SRS has focused on the total score only, and everyday EF measured with the BRIEF (the indexes BRI knowledge about how the different subscales are related to EF will and MI) in a clinical sample of children with ASD. We also provide new and more differentiated knowledge about the social investigated potential sex differences by comparing girls and deficiencies in ASD and may have important implications for boys, and possible age differences by splitting the sample at 12 interventions. years of age. We hypothesized that there is a significant positive Furthermore, it remains unclear what types of EF deficits are association between parental reports of SRS scores and BRIEF most relevant for the social problems of children and adolescents scores in children with ASD, controlling for age, IQ and sex. with ASD. It is known that the intelligence quotient (IQ), age Furthermore, we hypothesized that both BRI and MI from the and sex are factors that might influence the EF in children BRIEF are significant predictors of the SRS total score. with ASD, supported by findings of increasing EF deficits with age (Rosenthal et al., 2013), sex-dependent EF deficits (Bolte METHODS et al., 2011; Lemon et al., 2011; Lehnhardt et al., 2016)anda relationship between EF and intelligence (Diamond, 2013; Blijd- Participants Hoogewys et al., 2014; Rommelse et al., 2015). Thus, the degree The study was part of the national BUPgen network, recruiting to which IQ, sex, and age influence the relationship between EF patients from Norwegian health services specializing in the and social problems needs to be clarified. assessment of ASD and other neurodevelopmental disorders. There is an increased interest in possible sex differences in The current sample comprised 86 children with ASD, recruited ASD, since there might be important biological and behavioral between 2013 and 2016 and assessed at age 6–18 years. Thirteen differences between girls and boys with ASD (Halladay et al., of the children (15.1%) had childhood autism, one had atypical 2015; Lai et al., 2015). However, there are inconsistent findings autism (1.2%), 41 (47.7%) had Asperger syndrome and 31 (36%) regarding differences in the composition of EF difficulties in had unspecified pervasive developmental disorder (PDD-NOS) girls and boys with ASD. Some have found that girls have (Table 1). more EF difficulties with inhibition (Lemon et al., 2011), while Male: female ratio was 2.7:1. In total, 42 children were others reported that girls outperform boys on EF tasks related to diagnosed with at least one comorbid disorder. Attention processing speed and verbal fluency (Bolte et al., 2011; Lehnhardt deficit/hyperactivity disorder (ADHD) was the most common et al., 2016). White et al. (2017),showedthatgirlswithASDhave comorbid diagnosis, and 28 participants (32.6%) had an more problems related to everyday EF than boys. To the best of ADHD diagnosis in combination with their ASD diagnosis. our knowledge, there are no studies of possible sex differences in Furthermore, we divided our sample into; girls and boys, and the relationship between everyday EF deficits and social function two age groups above and below 12 years (6–12 years and 13– related to ASD symptoms. However, Bolte et al. (2011) found a 18 years). Our sample included 23 girls and 63 boys. There correlation between EF difficulties on performance based tasks were no significant differences between girls and boys on and stereotypical behaviors in boys. age, IQ or proportion of comorbid ADHD. More boys had a AreviewshowedthatEFcontinuestodevelopthroughout diagnosis of childhood autism, but there were no significant childhood in typical developed children, reaches adult-like levels differences between the sexes in the distribution of the Asperger in mid-adolescence, and that the different EF components vary in syndrome or PDD-NOS diagnoses. There were 36 children in their developmental trajectories (Best and Miller, 2010). Van den the youngest age group and 50 adolescents in the oldest group. Bergh et al. (2014) found that inhibition problems in everyday In these two age groups there were no significant differences life were more pronounced in young ASD children, and that in sex distribution, IQ, proportion of ADHD or ASD diagnoses planning was more evident for the oldest group. Contrary to (Table 1). All participants had an IQ within the normal range their expectations they did not find a relationship between ASD based on a standardized Wechsler’s test (Total IQ 70) and ≥ severity and EF (Van den Bergh et al., 2014). In a longitudinal spoke Norwegian fluently. The participants total IQ fell within 2 study by Andersen et al. (2015a) maturation of inhibition and standard deviations of the average normal score when including cognitive flexibility from childhood to adolescence was found in confidence intervals. Exclusion criteria were significant sensory ASD, even though the ASD group was more impaired in EF than losses (vision and/or hearing). typically developing (TD) children. They concluded that there may be a delayed development of EF in ASD, and suggested a Clinical Assessment possible developmental arrest for working memory (Andersen The children were assessed by a team of experienced clinicians et al., 2015b). Rosenthal et al. (2013) on the other hand showed (clinical psychologists and/or child psychiatrists). Diagnostic that there were widening divergences with age between children conclusions were best-estimate clinical diagnoses derived from with ASD and TC on EF tasks in everyday life, especially in tests, interview results and observations. All diagnoses were metacognitive executive abilities. They also reported significant based on the International Statistical Classification of Diseases and quite stable problems with flexibility in ASD (Rosenthal et al., and Related Health Problems 10th Revision (ICD-10) (World 2013). Health Organization, 1992)criteria,andtheautisticsymptoms

Frontiers in Behavioral Neuroscience | www.frontiersin.org 3 January 2018 | Volume 11 | Article 258 Torske et al. Executive and Social Function in ASD

TABLE 1 | Child characteristic.

Total (n 86) Girls (n 23) Boys (n 63) p-value 6–12 years (n 36) 13–18 years (n 50) p-value = = = = =

Girls/boys 23/63 23 63 11/25 12/38 0.622a Childhood autism (F84.0) 13 (15.1%) 1 (4.3%) 12 (19.0%) n/a 7 (19.4%) 6 (12.0%) 0.462a Atypical autism (F84.1) 1 (1.2%) 1 (4.3%) 0 n/a 1 (2.8%) 0 n/a Asperger syndrome (F84.5) 41 (47.7%) 12 (52.2%) 29 (47.7%) 0.142a 17 (47.2%) 24 (48.0%) 0.462a PDD-NOSb (F84.9) 31 (36.0%) 9 (39.1%) 22 (34.9%) 0.142a 11 (30.6%) 20 (40.0%) 0.462a Comorbid ADHDc 28 (32.6%) 6 (26.1%) 22 (34.9%) 0.604a 11 (30.6%) 17 (34.0%) 0.818a

Mean (SD) Range Mean (SD) Mean (SD)Mean(SD)Mean(SD)

Age 13.0 (2.7) 6–18 13.4 (2.4) 12.9 (2.9) 0.417 10.5 (1.7) 14.9 (1.7) <0.001* Total IQd 93.4 (14.5) 71–133 89.8 (10.8) 94.7 (15.5) 0.164 93.7 (12.0) 93.1 (16.2) 0.867 VerbalIQ 91.2(17.9) 58–134 87.4(15.7) 92.7(18.6) 0.231 89.9(15.0) 92.2(19.9) 0.555 Performance IQ 105.3 (16.3) 59–142 100.5 (11.2) 107.1 (17.5) 0.097 107.2 (15.0) 104.0 (17.1) 0.364

*p < 0.01. n/a, Not Applicable. aChi-square. bPDD-NOS, Pervasive developmental disorder unspecified. cADHD, Attention deficit/hyperactivity disorder. dIQ, Intelligence Quotient. were evaluated using the Autism Diagnostic Observation with a clinical diagnosis of Autistic Disorder, Asperger’s Disorder, Schedule (ADOS) (Lord et al., 2000) and/or Autism Diagnostic or more severe cases of pervasive developmental disorder not Interview-Revised (ADI-R) (Rutter et al., 2003b)and/orthe otherwise specified (PDD-NOS). T-scores of 60–75 represent Social Communication Questionnaire (SCQ) (Rutter et al., mild to moderate deficits in reciprocal social behavior that is 2003a). In addition, the assessment included a full medical clinically significant, resulting in mild to moderate interference and developmental history, physical examination, and IQ in everyday social interactions. assessment. The current study included children with ASD and comorbid ADHD. Results from recent studies indicate Executive Function (EF) that neuropsychiatric disorders overlap with respect to both In order to assess EF the parents completed the parent version symptoms and causes (Moreno-De-Luca et al., 2013), and it has of the BRIEF (Gioia et al., 2000). The BRIEF for children and been recognized by both clinicians and researchers for some time adolescents aged 5–18 years includes 86-item parent and teacher that ASD and ADHD often co-occur (Yerys et al., 2009). In the forms that allow professionals to assess everyday EFs in the DSM-5, other behavioral diagnosis may accompany a diagnosis home and school environments (Gioia et al., 2000). The BRIEF of ASD, for example ADHD (American Psychiatric Association, contains eight clinical scales that are grouped in a Behavioral 2013). Regulation Index (BRI): Inhibit, Shift and Emotional Control, and a Metacognition Index (MI): Initiate, Working Memory, Measures Plan/Organize, Organization of Materials and Monitor. T-scores Social Function of 65 are considered to represent clinically significant areas. ≥ The parent version of the SRS (developed for the age group The Global Executive Composite (GEC), is a summary score that 4–18 years) (Constantino and Gruber, 2005)wasusedto incorporates all eight clinical scales. The GEC has high reliability in both standardized and clinical samples (Cronbach’s alpha identify social communication difficulties. The SRS is used in = screening and/or as an aid to a clinical diagnosis of ASD and 0.80–0.98). The current study used the Norwegian version of is comprised of 65 questions, rated on a 4-point Likert scale. the parent rating form, which has high internal consistency (Cronbach’s alpha 0.76–0.92) (Fallmyr and Egeland, 2011). In addition to a total score, the SRS consists of five subscales: = Social Awareness, Social Cognition, Social Communication, Similar levels are reported for the English version (Cronbach’s alpha 0.80–0.98) (Gioia et al., 2000). Social Motivation and Autistic Mannerisms. The SRS has been = translated into Norwegian (Ørbeck, 2009). Internal consistency is high, both in population based samples and clinical samples Intelligence Quotient (IQ) (Cronbach’s alpha 0.93–0.97) (Constantino and Gruber, 2005). IQ was assessed using age-appropriate Wechsler tests of = The SRS’s reliability and validity have proven satisfactory in intelligence (Wechsler, 2002, 2003, 2008). both population based and clinical samples in Europe and in USA (Bolte et al., 2008; Wigham et al., 2012), and correlate Statistical Analyses well with Autism Diagnostic Interview-Revised (ADI-R) scores Data analyses were conducted using the statistical package (Constantino et al., 2003). T-scores of 76 are strongly associated IBM SPSS Statistics for Windows, version 23.0 (SPSS, Inc., ≥

Frontiers in Behavioral Neuroscience | www.frontiersin.org 4 January 2018 | Volume 11 | Article 258 Torske et al. Executive and Social Function in ASD

Chicago, Illinois). Descriptive analyses and bivariate correlations and Plan/Organize (p 0.035). There were no significant = were conducted. Pearson’s independent t-test was used to differences between the two age groups on BRIEF and SRS scores. compare means between girls and boys, and the two age However, there was a tendency for the youngest group to have groups. Chi-square for crosstabs was used to investigate higher scores (more problems) on the Social Awareness subscale differences in the distribution of autism diagnoses and comorbid from the SRS (p 0.036). = ADHD. The differences between the subgroups correlation coefficients were calculated using http://vassarstats.net/rdiff. html (two-tailed). Because of small subgroup sizes and no The Relationship between the SRS and the significant differences between the correlation coefficients, the BRIEF regression analyses were done on the total sample and sex There was a statistically significant (p < 0.001–0.005) relationship and age were incorporated as independent variables. The between all the subscales on the SRS and the BRIEF indexes, assumptions of linearity, multicollinearity, independence of except the SRS subscale Social Motivation and the BRIEF index errors, homoscedasticity, unusual points and normality of BRI. The proportion of shared variance (r2)variedbetween6 residuals were met. Separate multiple regression analyses were and 37%. The strongest positive correlation was found between conducted to explain the variance in SRS scores (total and the SRS total score and the BRIEF index scores GEC (r 0.62, = subscales) from age, sex, total IQ and BRIEF (BRI and MI p < 0.001) and MI (r 0.60, p < 0.001). There was also a = subscales). Bonferroni correction was used to correct significance moderate positive correlation between the SRS total score and level for multiple comparisons on all the analyses in this paper, BRI (r 0.48, p < 0.001). For Pearson’s correlation coefficients, = and the p < 0.01 level (p < 0.05/5 p < 0.01) is used in all the see Table 3A. = regression analyses. The p < 0.003 level is used for the correlation Girls generally had a stronger relationship between the SRS analyses (p < 0.05/18 p < 0.003). For all the other analyses (t- total and the BRIEF index scores than boys (see Table 3B). = tests, Chi-square and difference between correlation coefficients) However, the differences between the correlation coefficients we used p < 0.01. for girls and boys were not significant, calculated using http:// vassarstats.net/rdiff.html (to-tailed) (p 0.197–0.390). For girls = Ethical Considerations there was a strong and significant relationship between the The study was approved by the Regional Ethical Committee and SRS total and all the BRIEF indexes. The proportion of shared the Norwegian Data Inspectorate (REK #2012/1967), and was variance (r2) for the correlations for girls was 41–58%. For conducted in accordance with the Helsinki Declaration of the boys there was a moderate to strong relationship between the World Medical Association Assembly. The study is based on SRS total and the BRIEF index scores, and all the correlations tests that are included in regular clinical assessment. The patient, were significant. The proportion of shared variance (r2)forthe the family and the professional network around them were correlations for boys was 24–34%. offered information about the test results, the diagnostic process For the youngest age group (6–12 years) there was a strong and recommended interventions. Written informed consent was relationship between the SRS total and the BRIEF index scores, obtained from all the individual participants included in the and all the relationships were significant. For the oldest group study. (13–18 years) the relationship between BRI and SRS total was not significant. However, the MI and GEC showed significant RESULTS relationships to the SRS total. The proportion of shared variance (r2) for the correlations was 45–61% for the youngest group SRS and BRIEF Scores and 12–23% for the oldest group (see Table 3B). The differences The mean SRS total score was in the severe range (T-score 78.5). between the correlation coefficients between the youngest and the = The highest mean score was on the Social Mannerisms subscale oldest group were not significant, but showed strong tendencies (T-score 80.3). The lowest mean score was on the Social in the relationship between SRS total and GEC (p 0.021), = = Awareness subscale (T-score 65.3). Means and standard between SRS total and MI (p 0.024) and SRS total and BRI = = deviations for SRS total score and treatment subscales are shown (p 0.044). = in Table 2. Amultipleregressionanalysiswasconductedtoidentify The results from the BRIEF were in the clinically significant the relations between SRS total score and age, sex, total IQ, range on the indexes BRI (T-score 68.5) and GEC (T- BRIEF BRI and BRIEF MI. This model statistically significantly = score 67.1). The mean MI score from the BRIEF was T- explained SRS total; F 12.57, p < 0.001, R2 0.440. = (5, 80) = = score 64.4. The highest mean score was found on the subscale Only BRIEF MI had a significant independent contribution to the = Shift (T-score 72.2), and the lowest mean score was found on prediction, p 0.001. This result remained when the children = = the subscale Organization of Materials (T-score 54.7). All other with comorbid ADHD were removed from the analysis. For = subscales had T-score averages between 62 and 67 (Table 2). the children with ASD without ADHD (n 58), the regression = There were no significant differences between girls and boys model with age, sex, total IQ, BRIEF BRI, and BRIEF MI on BRIEF and SRS scores. However, there was a strong tendency significantly explained SRS total; F 10.31, p < 0.001, (5, 52) = for girls to have higher scores (more problems) on SRS total R2 0.498. Only BRIEF MI had a significant independent = (p 0.013) and Social Cognition (p 0.013), and a tendency for contribution to the prediction, p 0.003. Regression analyses = = = higher scores on the subscales Social Communication (p 0.032) were also done with each SRS subscale as dependent variables =

Frontiers in Behavioral Neuroscience | www.frontiersin.org 5 January 2018 | Volume 11 | Article 258 Torske et al. Executive and Social Function in ASD

TABLE 2 | T-scores for the Social Responsiveness Scale (SRS) and the Behavior Rating Inventory of Executive Function (BRIEF).

Total (n 86) Girls (n 23) Boys (n 63) p-valuea 6–12 years 13–18 years p-valuea = = = Mean (SD)Mean(SD)Mean(SD)Mean(SD)Mean(SD)

SOCIAL RESPONSIVENESS SCALE (SRS) SRS total 78.5 (15.0) 85.1 (16.4) 76.1 (13.8) 0.013 78.7 (14.6) 78.4 (15.4) 0.929 Social awareness 65.3 (13.4) 68.9 (15.3) 64.0 (12.5) 0.135 68.9 (10.9) 62.8 (14.6) 0.036 Social cognition 74.3 (16.6) 81.6 (19.8) 71.6 (14.5) 0.013 73.1 (17.0) 75.1 (16.4) 0.591 Social communication 73.9 (13.8) 79.2 (14.9) 72.0 (13.0) 0.032 75.0 (13.5) 73.2 (14.1) 0.551 Social motivation 78.4 (13.0) 82.8 (10.3) 76.8 (13.6) 0.060 76.7 (12.5) 79.7 (13.3) 0.288 Social mannerisms 80.3 (17.5) 86.0 (19.4) 78.2 (16.4) 0.068 79.4 (18.3) 80.9 (17.1) 0.714 BEHAVIOR RATING INVENTORY OF EXECUTIVE FUNCTION (BRIEF) Global executive composite (GEC) 67.1 (11.3) 67.9 (10.1) 66.8 (11.8) 0.707 66.7 (12.7) 67.4 (10.4) 0.792 Behavioral regulation index (BRI) 68.5 (13.5) 66.1 (13.1) 69.3 (13.7) 0.328 68.0 (14.2) 68.8 (13.1) 0.801 Metacognition index (MI) 64.4 (10.3) 67.3 (9.0) 63.3 (10.6) 0.107 64.0 (11.6) 64.6 (9.4) 0.795 Inhibit 63.0 (14.5) 62.1 (15.2) 63.3 (14.4) 0.749 62.6 (13.9) 63.2 (15.1) 0.838 Shift 72.2 (13.9) 67.6 (12.8) 73.8 (13.9) 0.060 71.4 (14.5) 72.7 (13.5) 0.668 Emotional control 64.7 (12.7) 62.9 (13.2) 65.4 (12.5) 0.433 64.3 (13.8) 65.0 (11.9) 0.823 Initiate 62.8 (11.1) 63.8 (10.3) 62.4 (11.5) 0.598 60.7 (10.4) 64.2 (11.5) 0.150 Workingmemory 67.2(10.6) 68.6(10.0) 66.7(10.9) 0.460 67.2(10.4) 67.2(10.9) 0.982 Plan/organize 63.7 (11.0) 67.9 (10.4) 62.2 (11.0) 0.035 63.2 (12.6) 64.1 (9.9) 0.695 Organization of materials 54.7 (11.1) 58.5 (9.5) 53.4 (11.3) 0.057 54.3 (12.1) 55.0 (10.4) 0.778 Monitor 62.3 (12.2) 63.1 (12.3) 62.1 (12.2) 0.736 62.6 (12.8) 62.2 (11.9) 0.867

*p < 0.01. aIndependent t-tests were conducted for comparisons between girls and boys, and between the age groups 6–12 years and 13–18 years. Elevated SRS T-scores indicate a high degree of impairment. T-scores of 76 or higher are strongly associated with a clinical diagnosis of ASD. T-scores of 60–75 indicate deficiencies in reciprocal social behavior that are clinically significant and are resulting in mild to moderate interference in everyday social interactions. Elevated BRIEF T-scores indicate a higher degree of impairment, with T-scores of 65 and above considered to represent clinically significant areas.

TABLE 3A | Associations between social function (SRS) and executive function (BRIEF) assessed with questionnaires (N 86). = Social Responsiveness Scale (SRS)

SRS Social Social Social Social Autistic total awareness cognition communication motivation mannerisms

BRIEF Behavioral regulation index Pearson r 0.48* 0.33* 0.46* 0.47* 0.24 0.47* p-value <0.001 0.002 <0.001 <0.001 0.025 <0.001 BRIEF Metacognition index Pearson r 0.60* 0.36* 0.54* 0.55* 0.46* 0.57* p-value <0.001 0.001 <0.001 <0.001 <0.001 <0.001 BRIEF Global executive composite Pearson r 0.61* 0.39* 0.56* 0.57* 0.41* 0.58* p-value <0.001 <0.001 <0.001 <0.001 <0.001 <0.001

*Significance after correction for multiple testing is set to p < 0.003 (2-tailed). BRIEF, Behavior Rating Inventory of Executive Function. on the total sample (n 86), and all the regression models were DISCUSSION < = significant (p 0.001). For the subscales Social Communication, The Importance of the Metacognition Index Social Motivation and Social Mannerisms, only MI from The main finding of the present study was that the metacognitive the BRIEF had a significant independent contribution to component of EF (MI), was the most important factor the predictions. For the subscale Social Awareness, only age in explaining social function in children with ASD. We had a significant independent contribution to the prediction hypothesized that both BRI and MI from the BRIEF would (p 0.001). None of the independent variables made an = predict SRS scores in children with ASD. However, despite high independent contribution on the subscale Social Cognition, but BRI scores, the MI explained more of the social dysfunctions both sex and MI showed strong tendencies (p 0.012 and 0.016). = measured with SRS in children with ASD. This is an interesting The details are described in Table 4. finding, since other studies have shown that behavior regulation

Frontiers in Behavioral Neuroscience | www.frontiersin.org 6 January 2018 | Volume 11 | Article 258 Torske et al. Executive and Social Function in ASD

TABLE 3B | Associations between social function (SRS) and executive function (BRIEF) assessed with questionnaires for the subgroups girls and boys, and the 0.13 0.50 0.10 55.36 0.97 13.49 / / / / / age groups 6–12 years and 13–18 years. / 1.73 0.17 0.22 SRS total SRS total 1.69 − − − −

Girls Boys 6–12 years 13–18 years 0.0001* n 23 n 63 n 36 n 50 < = = = =

BRIEF Behavioral Pearson r 0.64* 0.49* 0.67* 0.34 0.80 0.089 0.04 0.630 − regulation index p-value 0.001 <0.001 <0.001 0.015 − BRIEF Metacognition Pearson r 0.72* 0.54* 0.77* 0.47* 0.61 0.61 0.24 0.067 0.18 55.64 26.83 0.065 1.12 0.62 0.001* 0.27 index p-value <0.001 <0.001 <0.001 0.001 13.39 7.50 0.015 1.51 / / / / / / BRIEF Global Pearson r 0.76* 0.58* 0.78* 0.48* 1.68 0.04 0.27 1.42 14.86 − − − executive composite −

p-value <0.001 <0.001 <0.001 <0.001 −

*Significance after correction for multiple testing is set to p < 0.003 (2-tailed). 0.0001* < SRS, Social Responsiveness Scale. BRIEF, Behavior Rating Inventory of Executive Function. 0.05 0.680 0.53 0.358 − − is closely linked to social function (Kenworthy et al., 2009), 0.06 1.24 0.24 0.29 0.080 0.88 0.68 0.002* 0.25 77.97 20.39 0.253 8.97 5.99 0.112 / / / / / and that metacognition competence is of more importance for /

school performance (Carretti et al., 2014). MI is composed of ure of the Social Responsiveness Scale as outcome. 0.30 0.62 0.29 3.08 − − − subdomains like initiating, working memory, organizing and − monitoring. Difficulties in these areas are probably easier to 0.0001* overlook by parents, teachers and clinicians in everyday life, < than difficulties with subdomains within behavior regulation 0.12 0.192 like; inhibition, flexibility, and emotional control. Therefore, it 0.03 0.842 − is important to highlight that MI is of importance for social − function. The current findings provide new knowledge about 0.05 0.31 0.511 0.14 − 0.81 0.53 0.004* 0.18 0.51 60.29 49.31 0.001 20.64 12.17 2.94 0.334 / / / / / the relationship between EF and the various domains of social / competence in children with ASD, using parent-rated measures. 0.21 1.83 − There are few studies regarding the relationship between EF −

in everyday life settings and social function in children with 0.0001* ASD. The current findings are in line with Leung et al. (2015), < who showed that the MI from the BRIEF plays a role for 0.04 0.677 social function in children with ASD. However, in contrast to 0.94 0.038 − the finding from Leung et al. (2015),ourstudydidnotfinda − statistically significant contribution of the BRI to social function. 0.18 0.92 0.48 0.006* 0.14 0.58 0.64 0.26 0.041 0.01 50.28 32.92 0.019 5.55 16.11 6.42 0.029 0.66 / / / / / There is no obvious reason for these different findings, and / 0.24 the participants in the two studies share many of the same 1.60 16.72 − − characteristics. However, some differences in the recruitment −

process or sample size may explain the different results, and 0.0001* further studies are needed to clarify this question. Our study < contained a larger number of participants and had a more 0.03 0.756 conservative significant level compared to Leungs study, and 0.51 0.354 − − a proportion of our participants also had comorbid ADHD, 86). = N and this may explain the different results. However, the results 0.65 0.72 0.51 0.016 0.10 0.42 0.33 0.032 0.03 0.25 − 67.35 16.7811.24 0.322 9.08 0.012 2.04 / / / / from both studies emphasize the importance of the relationships / / 0.11 0.11 2.54 between MI and social function in children and adolescents with 0.94 − − − ASD. Kenworthy et al. (2009),ontheotherhand,foundthat − 0.277 0.367 0.391 0.247 0.372 0.440 EF such as BRI predicted only the communication symptoms 0.0001* and not the reciprocal social interaction in children with ASD. < However, they used a composite score based on ADI and ADOS BSig.95%BSig.95%BSig.95%BSig.95%BSig.95%BSig.95% scores to measure social function, and not the SRS (Kenworthy 1.59 0.001* −

et al., 2009). Landa and Goldberg (2005) found no relation Regression models summary ( 2 between the neuropsychological assessments of EF and the social R function in ASD children. The difference to our findings may be 0.01 (2-tailed). < -value p BRIEF-MI 0.37 0.044 0.01 Model’s Total IQ 0.07 0.432 BRIEF-BRI 0.16 0.234 Age Sex 5.15 0.096 Model’s p a result of biases in parent rated approaches (e.g., an inclination TABLE 4 | Factors SocialConstant awareness 38.37 0.010 9.40 Social cognition Social communication* B, Unstandardized Beta coefficients; Social Sig., motivation significant level; 95%, confidence interval for B for each factor entered in the regression models with meas Social mannerism SRS Total

Frontiers in Behavioral Neuroscience | www.frontiersin.org 7 January 2018 | Volume 11 | Article 258 Torske et al. Executive and Social Function in ASD to score “favorable” or “unfavorable” to items, regardless of the different relationship between EF and IQ in individuals with specific content), but it is also reasonable to assume that a parent- ASD might be the role of speed of information processing. rated design may uncover some relevant aspects of everyday This ability can be intact in ASD, and not correlated to function not accessible to the controlled setting of standardized IQ. However, speed of information processing is normally assessments. closely linked to the general factor (g-factor) of intelligence We did not find any significant differences between girls and (Scheuffgen et al., 2000; Wallace et al., 2009). The g-factor was boys in our sample. However, girls had a tendency for higher established by Spearman, who discovered that most cognitive scores on especially SRS total and Social Cognition, which might tests are positively correlated with each other, regardless of which imply that girls have more social problems than boys in our cognitive domain the individual test measures (Colom et al., sample. Contrary to White et al. (2017), the girls in our sample 2006). A review by Sheppard and Vernon (2008) concluded did not have significantly higher scores on the BRIEF than the that measures of mental speed and information processing are boys. Others have found that females with ASD have more EF significantly correlated with measures of intelligence in non- impairments compared to males (Lemon et al., 2011), while clinical samples. This highlights the importance of incorporating Lehnhardt et al. (2016) and Bolte et al. (2011) found evidence for IQ as a covariant in the relationship between the EF and social higher EF functioning for females with ASD than for males. Sex function. did not have an independent impact in our regression models, Earlier research has shown that TD children have a different but showed strong tendencies toward significance on the SRS relationship between EF and social function than children with total score and the subscale Social Cognition, where girls had ASD. In TD children, the BRI has an impact on their social more problems than boys in our sample. None of the studies function (Leung et al., 2015). Furthermore, there is evidence earlier mentioned have investigated the relationship between EF suggesting that there are mutual interactions between EF and and social function in everyday life settings, but it underlines the social function for both TD and ASD groups (Moriguchi, 2014; importance of being aware of possible sex differences in ASD. Leung et al., 2015). From studies of TD children, we know that Even though we did not find any significant differences social interaction may facilitate the development of EF. Especially between the two age groups children (6–12 years) and adolescents maternal scaffolding, modeling and imitation have demonstrated (13–18 year), there were some interesting tendencies. The BRIEF to be predictors of the development of EF (Moriguchi, 2014). scores in our sample were quite similar in the two age groups, It is also likely that EF facilitates the development of cognitive which is in contrast to Rosenthal et al. (2013) who found greater skills that are important for social interaction (van Lier and EF problems especially in metacognition for older children with Deater-Deckard, 2016). Subcomponents of EF, such as inhibition, ASD. Our result, that the relationship between EF difficulties and flexibility and monitoring are thought to influence the ability social dysfunction was strongest in the youngest group, might to engage in positive social interactions (van Lier and Deater- be due to young children having more generalized difficulties Deckard, 2016). White (2013) offers a different interpretation of in both social function and EF than older children. However, the difficulties that individuals with ASD experience on especially we found that in the oldest group there was a significant unstructured EF tasks. She argues that poor results on these tasks relationship between social function and metacognition, but not are caused by difficulties in forming an implicit understanding of for social function and behavior regulation. This might imply that the experimenter’s expectations for the task. In this view, implicit metacognitive EF is of importance for social function in older age, information is less available to individuals with ASD due to their and that behavior regulation does not have the same impact on mentalizing difficulties, and this leads to poor performance, more social function. In our regression analyses we showed that age than difficulties with EF per se. significantly predicted the ability to perceive social cues (Social Awareness), were the youngest group had more problems than the oldest group. Potential Implication for Development of Intelligence (IQ level) did not have any significant impact Clinical Interventions on the relationship between EF and social function in our Our finding that subdomains of social function have a different analysis. The relationship between measurements of EF, and relation to EF may be relevant for stratifying the treatment for especially fluid IQ, has been the subject of a longstanding children with ASD. Interventions that target metacognitive skills discussion. Several studies have found that much of the variance have, in earlier studies, shown to improve the social abilities in EF performance can be explained by IQ level (Joseph and in children and adolescents with ASD (Kenworthy et al., 2014; Tager-Flusberg, 2004; Friedman et al., 2006; Diamond, 2013). Leung et al., 2015). However, as social ability is a broad concept However, there is indication that the relationship between EF comprising different abilities and motivational factors, more and IQ is different among individuals with ASD, compared studies are needed to identify which part of social function may with typically developed (TD) and other patients groups. profit on metacognitive interventions. Our study indicates that Merchan-Naranjo et al. (2016) found that EF was affected, but the ability to perceive social cues and social motivation has, did not correlate with IQ in children and adolescents with possibly, a weaker relationship to EF than the other aspects ASD without intellectual disability. Rommelse et al. (2015) of social functioning. This might imply that children with even found that participants with above average intelligence problems related to social communication, social mannerisms performed relatively more poorly on some EF tasks compared and social cognition might benefit more from intervention to IQ matched controls. One possible explanation for the programs designed to enhance EF in general, and metacognitive

Frontiers in Behavioral Neuroscience | www.frontiersin.org 8 January 2018 | Volume 11 | Article 258 Torske et al. Executive and Social Function in ASD function, in particular. Individuals that primarily have problems children with ASD (Joseph et al., 2005). Since our participants with the ability to pick up on social cues and social motivation were school-aged children within the normal IQ range, we might be less likely to benefit from such intervention programs. assume that our finding is valid, even without a comprehensive Taken together, it is possible to hypothesize that children with language assessment. The current sample included a wide age difficulties with understanding social cues (social awareness) range and we therefore controlled for this in the regression might benefit more from classic social skills training programs analyses. Despite this, the age range might be a methodological (Chang et al., 2014; Otero et al., 2015). For children with limitation due to the correlation with social awareness, and problems related to social motivation it might be important to future studies might benefit from focus on more restricted age build on the child’s areas of interest to enhance their motivation groups. (Dunst et al., 2011). In all cases, we need to assess the individual Our study did not clarify in which direction EF and child’s cognitive and social profile and then tailor interventions social function affect each other, but it suggests that there to fit the child. Further studies are needed to examine the is a strong relation between the two. The current findings clinical implications of these findings. More generally, the finding should be replicated in independent samples and combined that there is an important relationship between EF and social with objective measurements, such as neuropsychological or function, gives support to the hypothesis that an intervention neurophysiological examination. This was not available in that has an impact on cognitive abilities and EF is also likely the current sample. Thus, future research should combine to have an influence on the social skills of children and laboratory and informant-based measures for a more in-depth adolescents with ASD (McGovern and Sigman, 2005; Wang et al., investigation of the link between EF and social function, as 2017). the two measures are complementary to each other (Leung et al., 2015). To fully understand the relationship between Strengths and Limitations of the Study EF and social functioning, studies with longitudinal designs The strengths of the study include a reasonable sample size of are need to provide more specific detail about functional clinically well-defined children and adolescents with ASD, with developmental change at different ages (Taylor et al., 2015). amoderatesampleoffemales.BoththeSRSandtheBRIEF Furthermore, most research within the ASD field is conducted are parental ratings, and this might bias the findings. However, on male samples, but some evidence suggests that females exhibit we considered that parents observe their children when they greater cognitive impairments than their male counterparts are engaged in a range of everyday activities and that their (Lemon et al., 2011; White et al., 2017). In our analysis, we observations add valuable information. Particularly since it is found a tendency for girls to have higher SRS scores, and known that difficulties may not be observable in a laboratory a stronger relationship between social function and EF than setting or structured clinical settings, informant measures might boys. For future research, it is important to investigate if the have higher ecological validity (Kenworthy et al., 2008). At the relationship between EF and social function can be modulated same time, parent-ratings are likely to be affected by other factors by sex. such as IQ, emotional/behavioral problems and comorbidities in the children (Aldridge et al., 2012; Cholemkery et al., 2014; CONCLUSION Havdahl et al., 2016). However, diagnostic tools like the ADOS and the ADI-R are also influenced by these kinds of confounding We report a relationship between parental reports of EF and factors (Havdahl et al., 2016). Some of the children in our social function in an everyday setting in children with ASD. We study were medicated, and several had a comorbid disorder. found that the metacognitive domain of EF has a significant It is known from the literature that comorbidity is common association to many aspects of social function. These results may in ASD, and that as many as 70% have a comorbid disorder have implications for understanding the cognitive components (Simonoff et al., 2008; Lai et al., 2014; Gjevik et al., 2015). The in the social deficits that define ASD. Further studies are needed main finding that metacognitive aspects of EF are significantly to clarify if children with ASD will improve their social function related to social function was significant also after removing through intervention programs designed to enhance EF in the children with comorbid ADHD from the analysis. It can general and metacognitive function, in particular. be argued that it is important to study individuals with ASD and comorbid ADHD because this is a common comorbidity. A recent meta-analysis of EF and children and adolescents with AUTHOR CONTRIBUTIONS high-functioning ASD, was the first to take the confounding effect of ADHD comorbidity into account. They confirmed the Conceived and designed the study: TT, TN, MØ, and OA. presence of executive dysfunction in this group, and found that Performed the study: TT, TN, and NS. Analyzed the data: TT these deficits were not solely accounted for by the effect of and TN. Interpreted the data, wrote the paper, and approved the comorbid ADHD or general cognitive abilities (Lai et al., 2016). version to be published: TT, TN, MØ, NS, and OA. Furthermore, language deficits are important factors that affect social communication/function in a negative way in ASD. We did ACKNOWLEDGMENTS not perform a comprehensive assessment of different language functions in our participants, but some studies have found that We are thankful to the BUPgen partners and to all the EF is not related to language ability in verbal, school-aged participants. The study is part of the BUPgen Study group and

Frontiers in Behavioral Neuroscience | www.frontiersin.org 9 January 2018 | Volume 11 | Article 258 Torske et al. Executive and Social Function in ASD the research network NeuroDevelop. The project was supported the Regional Research Network NeuroDevelop (Grant #39763). by the National Research Council of Norway (Grant #213694) The corresponding author has a research grant from Vestre Viken and the South-Eastern Norway Regional Health Authority funds Hospital Trust (Grant #6903002).

REFERENCES Constantino, J. N., Gruber, C. P., Davis, S., Hayes, S., Passanante, N., and Przybeck, T. (2004). The factor structure of autistic traits. J. Child Psychol. Psychiatry 45, Achenbach, T. M. (2011). Commentary: definitely more than measurement error: 719–726. doi: 10.1111/j.1469-7610.2004.00266.x but how should we understand and deal with informant discrepancies? J. Clin. Corbett, B. A., Constantine, L. J., Hendren, R., Rocke, D., and Ozonoff, S. (2009). Child Adolesc. Psychol. 40, 80–86. doi: 10.1080/15374416.2011.533416 Examining executive functioning in children with autism spectrum disorder, Aldridge, F. J., Gibbs, V. M., Schmidhofer, K., and Williams, M. (2012). attention deficit hyperactivity disorder and typical development. Psychiatry Res. Investigating the clinical usefulness of the social responsiveness scale (SRS) in 166, 210–222. doi: 10.1016/j.psychres.2008.02.005 atertiarylevel,autismspectrumdisorderspecificassessmentclinic.J. Autism Craig, F., Margari, F., Legrottaglie, A. R., Palumbi, R., De Giambattista, C., Dev. Disord. 42, 294–300. doi: 10.1007/s10803-011-1242-9 and Margari, L. (2016). A review of executive function deficits in autism American Psychiatric Association (2013). Diagnostic and Statistical spectrum disorder and attention-deficit/hyperactivity disorder. Neuropsychiatr. Manual of Mental Disorders (DSM-5). Washington, DC: APA. Dis. Treat. 12, 1191–1202. doi: 10.2147/NDT.S104620 doi: 10.1176/appi.books.9780890425596 Dawson, G., Munson, J., Estes, A., Osterling, J., McPartland, J., Toth, K., et al. Andersen, P. N., Skogli, E. W., Hovik, K. T., Egeland, J., and Oie, (2002). Neurocognitive function and joint attention ability in young children M. (2015a). Associations among symptoms of autism, symptoms of with autism spectrum disorder versus developmental delay. Child Dev. 73, depression and executive functions in children with high-functioning 345–358. doi: 10.1111/1467-8624.00411 autism: a 2 year follow-up study. J. Autism Dev. Disord. 45, 2497–2507. de Vries, M., and Geurts, H. (2015). Influence of autism traits and executive doi: 10.1007/s10803-015-2415-8 functioning on quality of life in children with an autism spectrum disorder. Andersen, P. N., Skogli, E. W., Hovik, K. T., Geurts, H., Egeland, J., J. Autism Dev. Disord. 45, 2734–2743. doi: 10.1007/s10803-015-2438-1 and Oie, M. (2015b). Working memory arrest in children with high- Diamond, A. (2013). Executive functions. Annu. Rev. Psychol. 64, 135–168. functioning autism compared to children with attention-deficit/hyperactivity doi: 10.1146/annurev-psych-113011-143750 disorder: results from a 2-year longitudinal study. Autism 19, 443–450. Dunst, C. J., Trivette, C. M., and Masiello, T. (2011). Exploratory investigation doi: 10.1177/1362361314524844 of the effects of interest-based learning on the development of young Barendse, E. M., Hendriks, M. P., Jansen, J. F., Backes, W. H., children with autism. Autism 15, 295–305. doi: 10.1177/13623613103 Hofman, P. A., Thoonen, G., et al. (2013). Working memory deficits 70971 in high-functioning adolescents with autism spectrum disorders: Fallmyr, Ø., and Egeland, J. (2011). Psykometriske egenskaper for den norske neuropsychological and neuroimaging correlates. J. Neurodev. Disord. versjonen av behavior rating inventory of executive function (BRIEF). 5:14. doi: 10.1186/1866-1955-5-14 Tidsskrift Norsk Psykologforening 48, 339–343. Available online at: http://www. Best, J. R., and Miller, P. H. (2010). A developmental perspective on executive psykologtidsskriftet.no/index.php?seks_id=138594&a=2 function. Child Dev. 81, 1641–1660. doi: 10.1111/j.1467-8624.2010.01499.x Friedman, N. P., Miyake, A., Corley, R. P., Young, S. E., Defries, J. C., and Hewitt, Blijd-Hoogewys, E. M., Bezemer, M. L., and Van Geert, P. L. (2014). Executive J. K. (2006). Not all executive functions are related to intelligence. Psychol. Sci. functioning in children with ASD: an analysis of the BRIEF. J. Autism Dev. 17, 172–179. doi: 10.1111/j.1467-9280.2006.01681.x Disord. 44, 3089–3100. doi: 10.1007/s10803-014-2176-9 Geurts, H., De Vries, M., and Van Den Bergh, S. F. (2014a). “Executive functioning Bolte, S., Duketis, E., Poustka, F., and Holtmann, M. (2011). Sex differences in theory and autism,” in Handbook of Executive Functioning, eds S. Goldstein cognitive domains and their clinical correlates in higher-functioning autism and J. A. Naglieri (New York, NY: Springer Science and Business Media), spectrum disorders. Autism 15, 497–511. doi: 10.1177/1362361310391116 121–141. Bolte, S., Poustka, F., and Constantino, J. N. (2008). Assessing autistic traits: cross- Geurts, H. M., Van Den Bergh, S. F., and Ruzzano, L. (2014b). Prepotent response cultural validation of the social responsiveness scale (SRS). Autism Res. 1, inhibition and interference control in autism spectrum disorders: two meta- 354–363. doi: 10.1002/aur.49 analyses. Autism Res. 7, 407–420. doi: 10.1002/aur.1369 Carretti, B., Caldarola, N., Tencati, C., and Cornoldi, C. (2014). Improving reading Gilotty, L., Kenworthy, L., Sirian, L., Black, D. O., and Wagner, A. E. (2002). comprehension in reading and listening settings: the effect of two training Adaptive skills and executive function in autism spectrum disorders. Child programmes focusing on metacognition and working memory. Br. J. Educ. Neuropsychol. 8, 241–248. doi: 10.1076/chin.8.4.241.13504 Psychol. 84, 194–210. doi: 10.1111/bjep.12022 Gioia, G. A., Isquith, P. K., Guy, S. C., and Kenworthy, L. (2000). Behavior Chang, Y. C., Laugeson, E. A., Gantman, A., Ellingsen, R., Frankel, F., and Rating Inventory of Executive Function (BRIEF). Odessa, FL: PAR Psychological Dillon, A. R. (2014). Predicting treatment success in social skills training Assessment Resources, Inc. for adolescents with autism spectrum disorders: the UCLA program for Gjevik, E., Sandstad, B., Andreassen, O. A., Myhre, A. M., and Sponheim, E. the education and enrichment of relational skills. Autism 18, 467–470. (2015). Exploring the agreement between questionnaire information and DSM- doi: 10.1177/1362361313478995 IV diagnoses of comorbid psychopathology in children with autism spectrum Cholemkery, H., Kitzerow, J., Rohrmann, S., and Freitag, C. M. (2014). Validity disorders. Autism 19, 433–442. doi: 10.1177/1362361314526003 of the social responsiveness scale to differentiate between autism spectrum Halladay, A. K., Bishop, S., Constantino, J. N., Daniels, A. M., Koenig, K., Palmer, disorders and disruptive behaviour disorders. Eur. Child Adolesc. Psychiatry 23, K., et al. (2015). Sex and gender differences in autism spectrum disorder: 81–93. doi: 10.1007/s00787-013-0427-5 summarizing evidence gaps and identifying emerging areas of priority. Mol. Colom, R., Jung, R. E., and Haier, R. J. (2006). Distributed brain sites for the g- Autism 6:36. doi: 10.1186/s13229-015-0019-y factor of intelligence. Neuroimage 31, 1359–1365. doi: 10.1016/j.neuroimage. Havdahl, K. A., Hus Bal, V., Huerta, M., Pickles, A., Oyen, A. S., Stoltenberg, 2006.01.006 C., et al. (2016). Multidimensional influences on autism symptom measures: Constantino, J. N., Davis, S. A., Todd, R. D., Schindler, M. K., Gross, M. implications for use in etiological research. J. Am. Acad. Child Adolesc. M., Brophy, S. L., et al. (2003). Validation of a brief quantitative measure Psychiatry 55, 1054–1063.e3. doi: 10.1016/j.jaac.2016.09.490 of autistic traits: comparison of the social responsiveness scale with the Hill, E. L. (2004). Executive dysfunction in autism. Trends Cogn. Sci. 8, 26–32. autism diagnostic interview-revised. J. Autism Dev. Disord. 33, 427–433. doi: 10.1016/j.tics.2003.11.003 doi: 10.1023/A:1025014929212 Joseph, R. M., McGrath, L. M., and Tager-Flusberg, H. (2005). Executive Constantino, J. N., and Gruber, C. P. (2005). Social Responsiveness Scale (SRS) dysfunction and its relation to language ability in verbal school-age children Manual.LosAngeles,CA:WesternPsychologicalService. with autism. Dev. Neuropsychol. 27, 361–378. doi: 10.1207/s15326942dn2703_4

Frontiers in Behavioral Neuroscience | www.frontiersin.org 10 January 2018 | Volume 11 | Article 258 Torske et al. Executive and Social Function in ASD

Joseph, R. M., and Tager-Flusberg, H. (2004). The relationship of theory of mind Ørbeck, B. (2009). Sosial Responsivitets Skala. Norsk Oversettelse. Oslo: Hogrefe and executive functions to symptom type and severity in children with autism. Psykologiske Forlag. Dev. Psychopathol. 16, 137–155. doi: 10.1017/S095457940404444X Otero, T. L., Schatz, R. B., Merrill, A. C., and Bellini, S. (2015). Social skills training Kenworthy, L., Anthony, L. G., Naiman, D. Q., Cannon, L., Wills, M. C., Luong- for youth with autism spectrum disorders: a follow-up. Child Adolesc. Psychiatr. Tran, C., et al. (2014). Randomized controlled effectiveness trial of executive Clin. N. Am. 24, 99–115. doi: 10.1016/j.chc.2014.09.002 function intervention for children on the autism spectrum. J. Child Psychol. Pellicano, E. (2010). Individual differences in executive function and central Psychiatry 55, 374–383. doi: 10.1111/jcpp.12161 coherence predict developmental changes in theory of mind in autism. Dev. Kenworthy, L., Black, D. O., Harrison, B., Della Rosa, A., and Wallace, Psychol. 46, 530–544. doi: 10.1037/a0018287 G. L. (2009). Are executive control functions related to autism Pennington, B. F., and Ozonoff,S.(1996).Executivefunctionsand symptoms in high-functioning children? Child Neuropsychol. 15, 425–440. developmental psychopathology. J. Child Psychol. Psychiatry 37, 51–87. doi: 10.1080/09297040802646983 doi: 10.1111/j.1469-7610.1996.tb01380.x Kenworthy, L., Yerys, B. E., Anthony, L. G., and Wallace, G. L. (2008). Rommelse, N., Langerak, I., Van Der Meer, J., De Bruijn, Y., Staal, W., Understanding executive control in autism spectrum disorders in the lab and in Oerlemans, A., et al. (2015). Intelligence may moderate the cognitive profile the real world. Neuropsychol. Rev. 18, 320–338. doi: 10.1007/s11065-008-9077-7 of patients with ASD. PLoS ONE 10:e0138698. doi: 10.1371/journal.pone. Lai, C. L., Lau, Z., Lui, S. S., Lok, E., Tam, V., Chan, Q., et al. (2016). Meta- 0138698 analysis of neuropsychological measures of executive functioning in children Rosenthal, M., Wallace, G. L., Lawson, R., Wills, M. C., Dixon, E., Yerys, B. and adolescents with high-functioning autism spectrum disorder. Autism Res. E., et al. (2013). Impairments in real-world executive function increase from 10, 911–939. doi: 10.1002/aur.1723 childhood to adolescence in autism spectrum disorders. Neuropsychology 27, Lai, M. C., Lombardo, M. V., Auyeung, B., Chakrabarti, B., and Baron-Cohen, S. 13–18. doi: 10.1037/a0031299 (2015). Sex/gender differences and autism: setting the scene for future research. Rutter, M., Bailey, A., and Lord, C. (2003a). The Social Communication J. Am. Acad. Child Adolesc. Psychiatry 54, 11–24. doi: 10.1016/j.jaac.2014. Questionnaire (SCQ). Los Angeles, CA: Western Psychological Services. 10.003 Rutter, M., Lord, C., and Lecouteur, A. (2003b). Autism Diagnostic Interview- Lai, M. C., Lombardo, M. V., and Baron-Cohen, S. (2014). Autism. Lancet 383, Revised Manual (ADI-R). Los Angeles, CA: Western Psychological Services. 896–910. doi: 10.1016/S0140-6736(13)61539-1 Scheuffgen, K., Happe, F., Anderson, M., and Frith, U. (2000). High “intelligence,” Landa, R. J., and Goldberg, M. C. (2005). Language, social, and executive functions low “IQ”? Speed of processing and measured IQ in children with autism. Dev. in high functioning autism: a continuum of performance. J. Autism Dev. Disord. Psychopathol. 12, 83–90. doi: 10.1017/S095457940000105X 35, 557–573. doi: 10.1007/s10803-005-0001-1 Sheppard, L. D., and Vernon, P. A. (2008). Intelligence and speed of information- Lehnhardt, F. G., Falter, C. M., Gawronski, A., Pfeiffer, K., Tepest, R., processing: a review of 50 years of research. Pers. Individ. Dif. 44, 535–551. Franklin, J., et al. (2016). Sex-related cognitive profile in autism spectrum doi: 10.1016/j.paid.2007.09.015 disorders diagnosed late in life: implications for the female autistic Simonoff,E.,Pickles,A.,Charman,T.,Chandler,S.,Loucas,T.,and phenotype. J. Autism Dev. Disord. 46, 139–154. doi: 10.1007/s10803-015- Baird, G. (2008). Psychiatric disorders in children with autism spectrum 2558-7 disorders: prevalence, comorbidity, and associated factors in a population- Lemon, J. M., Gargaro, B., Enticott, P. G., and Rinehart, N. J. (2011). Executive derived sample. J. Am. Acad. Child Adolesc. Psychiatry 47, 921–929. functioning in autism spectrum disorders: a gender comparison of response doi: 10.1097/CHI.0b013e318179964f inhibition. J. Autism Dev. Disord. 41, 352–356. doi: 10.1007/s10803-010- Taylor, S. J., Barker, L. A., Heavey, L., and McHale, S. (2015). The longitudinal 1039-2 development of social and executive functions in late adolescence and early Leung, R. C., Vogan, V. M., Powell, T. L., Anagnostou, E., and adulthood. Front. Behav. Neurosci. 9:252. doi: 10.3389/fnbeh.2015.00252 Taylor, M. J. (2015). The role of executive functions in social Van den Bergh, S. F., Scheeren, A. M., Begeer, S., Koot, H. M., and Geurts, H. M. impairment in autism spectrum disorder. Child Neuropsychol. 22, 336–344. (2014). Age related differences of executive functioning problems in everyday doi: 10.1080/09297049.2015.1005066 life of children and adolescents in the autism spectrum. J. Autism Dev. Disord. Leung, R. C., and Zakzanis, K. K. (2014). Brief report: cognitive flexibility in 44, 1959–1971. doi: 10.1007/s10803-014-2071-4 autism spectrum disorders: a quantitative review. J. Autism Dev. Disord. 44, Van Eylen, L., Boets, B., Steyaert, J., Wagemans, J., and Noens, I. (2015). Executive 2628–2645. doi: 10.1007/s10803-014-2136-4 functioning in autism spectrum disorders: influence of task and sample Lord, C., Risi, S., Lambrecht, L., Cook, E. H. Jr., Leventhal, B. L., DiLavore, P. L., characteristics and relation to symptom severity. Eur. Child Adolesc. Psychiatry et al. (2000). The autism diagnostic observation schedule-generic: a standard 24, 1399–1417. doi: 10.1007/s00787-015-0689-1 measure of social and communication deficits associated with the spectrum of van Lier, P. A., and Deater-Deckard, K. (2016). Children’s elementary school social autism. J. Autism Dev. Disord. 30, 205–223. doi: 10.1023/A:1005592401947 experience and executive functions development: introduction to a special McGovern, C. W., and Sigman, M. (2005). Continuity and change from early section. J. Abnorm. Child Psychol. 44, 1–6. doi: 10.1007/s10802-015-0113-9 childhood to adolescence in autism. J. Child Psychol. Psychiatry 46, 401–408. Wallace, G. L., Anderson, M., and Happe, F. (2009). Brief report: doi: 10.1111/j.1469-7610.2004.00361.x information processing speed is intact in autism but not correlated Merchan-Naranjo, J., Boada, L., Del Rey-Mejias, A., Mayoral, M., Llorente, C., with measured intelligence. J. Autism Dev. Disord. 39, 809–814. Arango, C., et al. (2016). Executive function is affected in autism spectrum doi: 10.1007/s10803-008-0684-1 disorder, but does not correlate with intelligence. Rev. Psiquiatr. Salud Ment. Wang, Y., Zhang, Y. B., Liu, L. L., Cui, J. F., Wang, J., Shum, D. H., et al. (2017). A 9, 39–50. doi: 10.1016/j.rpsmen.2016.01.001 meta-analysis of working memory impairments in autism spectrum disorders. Miyake, A., Friedman, N. P., Emerson, M. J., Witzki, A. H., Howerter, A., and Neuropsychol. Rev. 27, 46–61. doi: 10.1007/s11065-016-9336-y Wager, T. D. (2000). The unity and diversity of executive functions and their Wechsler, D. (2002). Wechsler Preschool and Primary Scale of Intelligence, 3rd Edn. contributions to complex “Frontal Lobe” tasks: a latent variable analysis. Cogn. San Antonio, TX: Pearson. Psychol. 41, 49–100. doi: 10.1006/cogp.1999.0734 Wechsler, D. (2003). Wechler Intelligence Scale for Children, 4th Edn. San Antonio, Moreno-De-Luca, A., Myers, S. M., Challman, T. D., Moreno-De-Luca, D., Evans, TX: Pearson. D. W., and Ledbetter, D. H. (2013). Developmental brain dysfunction: revival Wechsler, D. (2008). Wechsler Adult Intelligence Scale, 4th Edn. San Antonio, TX: and expansion of old concepts based on new genetic evidence. Lancet Neurol. Pearson. 12, 406–414. doi: 10.1016/S1474-4422(13)70011-5 White, E. I., Wallace, G. L., Bascom, J., Armour, A. C., Register-Brown, K., Popal, Moriguchi, Y. (2014). The early development of executive function and H. S., et al. (2017). Sex differences in parent-reported executive functioning and its relation to social interaction: a brief review. Front. Psychol. 5:388. adaptive behavior in children and young adults with autism spectrum disorder. doi: 10.3389/fpsyg.2014.00388 Autism Res. 10, 1653–1662. doi: 10.1002/aur.1811 Nelson, A. T., Lopata, C., Volker, M. A., Thomeer, M. L., Toomey, J. A., and Dua, White, S. J. (2013). The triple I hypothesis: taking another(’s) perspective E. (2016). Exploratory factor analysis of SRS-2 teacher ratings for youth with on executive dysfunction in autism. J. Autism Dev. Disord. 43, 114–121. ASD. J. Autism Dev. Disord. 46, 2905–2915. doi: 10.1007/s10803-016-2822-5 doi: 10.1007/s10803-012-1550-8

Frontiers in Behavioral Neuroscience | www.frontiersin.org 11 January 2018 | Volume 11 | Article 258 Torske et al. Executive and Social Function in ASD

Wigham, S., McConachie, H., Tandos, J., Le Couteur, A. S., and Gateshead Conflict of Interest Statement: The authors declare that the research was Millennium Study Core, T. (2012). The reliability and validity of the social conducted in the absence of any commercial or financial relationships that could responsiveness scale in a UK general child population. Res. Dev. Disabil. 33, be construed as a potential conflict of interest. 944–950. doi: 10.1016/j.ridd.2011.12.017 World Health Organization (1992). The ICD-10 Classification of Mental and Copyright © 2018 Torske, Nærland, Øie, Stenberg and Andreassen. This is an open- Behavioral Disorders: Clinical Descriptions and Diagnostic Guidelines.Geneva: access article distributed under the terms of the Creative Commons Attribution World Health Organization. License (CC BY). The use, distribution or reproduction in other forums is permitted, Yerys, B. E., Wallace, G. L., Sokoloff,J.L.,Shook,D.A.,James,J.D.,and provided the original author(s) or licensor are credited and that the original Kenworthy, L. (2009). Attention deficit/hyperactivity disorder symptoms publication in this journal is cited, in accordance with accepted academic practice. moderate cognition and behavior in children with autism spectrum disorders. No use, distribution or reproduction is permitted which does not comply with these Autism Res. 2, 322–333. doi: 10.1002/aur.103 terms.

Frontiers in Behavioral Neuroscience | www.frontiersin.org 12 January 2018 | Volume 11 | Article 258 III

RESEARCH ARTICLE Autism spectrum disorder polygenic scores are associated with every day executive function in children admitted for clinical assessment Tonje Torske , Terje Nærland, Francesco Bettella, Thomas Bjella, Eva Malt, Anne Lise Høyland, Nina Stenberg, Merete Glenne Øie, and Ole A. Andreassen

Autism spectrum disorder (ASD) and other neurodevelopmental disorders (NDs) are behaviorally defined disorders with overlapping clinical features that are often associated with higher-order cognitive dysfunction, particularly executive dys- function. Our aim was to determine if the polygenic score (PGS) for ASD is associated with parent-reported executive dys- function in everyday life using the Behavior Rating Inventory of Executive Function (BRIEF). Furthermore, we investigated if PGS for general intelligence (INT) and attention deficit/hyperactivity disorder (ADHD) also correlate with BRIEF. We included 176 children, adolescents and young adults aged 5–22 years with full-scale intelligence quotient (IQ) above 70. All were admitted for clinical assessment of ASD symptoms and 68% obtained an ASD diagnosis. We found a significant difference between low and high ASD PGS groups in the BRIEF behavior regulation index (BRI) (P = 0.015, Cohen’s d = 0.69). A linear regression model accounting for age, sex, full-scale IQ, Social Responsiveness Scale (SRS) total score, ASD, ADHD and INT PGS groups as well as genetic principal components, significantly predicted the BRI score; F(11,130) = 8.142, P < 0.001, R2 = 0.41 (unadjusted). Only SRS total (P < 0.001), ASD PGS 0.1 group (P = 0.018), and sex (P = 0.022) made a significant contribution to the model. This suggests that the common ASD risk gene variants have a stronger association to behavioral regulation aspects of executive dysfunction than ADHD risk or INT variants in a clini- cal sample with ASD symptoms. Autism Res 2019, 00: 1–14. © 2019 The Authors. Autism Research published by Interna- tional Society for Autism Research published by Wiley Periodicals, Inc.

Lay Summary: People with autism spectrum disorder (ASD) often have difficulties with higher-order cognitive processes that regulate thoughts and actions during goal-directed behavior, also known as executive function (EF). We studied the association between genetics related to ASD and EF and found a relation between high polygenic score (PGS) for ASD and difficulties with behavior regulation aspects of EF in children and adolescents under assessment for ASD. Furthermore, high PGS for general intelligence was related to social problems.

Keywords: autism spectrum disorder; polygenic score; executive function; attention deficit/hyperactivity disorder; behavior rating inventory of executive function

Introduction risk of ASD have been identified, and these rare genetic vari- ants, for example, copy-number variations (CNVs), are esti- Autism spectrum disorder (ASD) is a neurodevelopmental mated to explain about 5–10% of the genetic risk for ASD disorder (ND) characterized by deficits in social communica- [Ramaswami & Geschwind, 2018]. However, most of the tion and interaction, and restricted and repetitive behaviors heritability of ASD seems to be caused by common genetic and interests [American Psychiatric Association, 2013]. The variants [Gaugler et al., 2014]. ASD, like attention defi- heritability of ASD is high, with estimates ranging from cit/hyperactivity disorder (ADHD), schizophrenia, and bipo- 70 to 90%, and the disorder is regarded as a complex genetic lar disorder, have been shown to be polygenic disorders, in disorder with multifactorial causes [Lai, Lombardo, & Baron- which each single risk variant has a small effect on the dis- Cohen, 2014]. Several rare geneticvariantsconferringhigh ease phenotype [Demontis et al., 2018; Grove et al., 2019;

From the Division of Mental Health and Addiction, Vestre Viken Hospital Trust, Drammen, Norway (T.T.); Department of Psychology, University of Oslo, Oslo, Norway (T.T., M.G.Ø.); NevSom, Department of Rare Disorders, Oslo University Hospital, Oslo, Norway (T.N.); NORMENT Centre, University of Oslo, Oslo, Norway (T.N., F.B., T.B., O.A.A.); Division of Mental Health and Addiction, Oslo University Hospital, Oslo, Norway (F.B., T.B., N.S., O.A.A.); Institute of Clinical Medicine, Ahus Campus, University of Oslo, Oslo, Norway (E.M.); Department of Adult Habilitation, Akershus University Hospital, Lørenskog, Norway (E.M.); Regional Centre for Child and Youth Mental Health and Child Welfare, Department of Mental Health, Faculty of Medicine and Health Sciences, Norwegian University of Science and Technology, Trondheim, Norway (A.L.H.); Department of Pediatrics, St. Olav Hospital, Trond- heim University Hospital, Trondheim, Norway (A.L.H.); Research Department, Innlandet Hospital Trust, Lillehammer, Norway (M.G.Ø.) Received February 12, 2019; accepted for publication August 24, 2019 Address for correspondence and reprints: Tonje Torske, Division of Mental Health and Addiction, Vestre Viken Hospital Trust, Drammen, Norway. E-mail: [email protected] This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribu- tion in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. Published online 00 Month 2019 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/aur.2207 © 2019 The Authors. Autism Research published by International Society for Autism Research published by Wiley Periodicals, Inc.

INSAR Autism Research 000: 1–14, 2019 1 International Schizophrenia Consortium et al., 2009; diagnosis might still have profound social difficulties and Schizophrenia Working Group of the Psychiatric Genomics executive dysfunction [American Psychiatric Association, Consortium, 2014]. Based on genome-wide association stud- 2013]. Comorbidity is common in ASD, and about 30% ies (GWASs) summary statistics for a given complex disor- also meet the diagnostic criteria for ADHD [Lord, Elsabbagh, der, it is possible to derive measures of the cumulative Baird, & Veenstra-Vanderweele, 2018]. Cognitive difficulties genetic loads for the disorders inherent to an individual’s are also observed in ADHD [Craig et al., 2016], and particu- genotype. Such measures are often called polygenic scores larly executive dysfunctions are present across diagnostic (PGS) [International Schizophrenia Consortium et al., categories like ASD and ADHD as well as many other psy- 2009]. Recently, Grove et al. identified five risk loci for chiatric disorders [Dajani, Llabre, Nebel, Mostofsky, & ASD based on a genome-wide association meta-analysis of Uddin, 2016; Snyder, Miyake, & Hankin, 2015]. Executive 18,381 cases of ASD and 27,969 controls [Grove et al., function is a heritable cognitive domain and deficits are 2019]. Even though these PGS are getting better at separat- found in unaffected ASD family members at higher rates ing ASD cases from controls [Grove et al., 2019], their sensi- than in the general population [Benca et al., 2017]. Further- tivity and specificity are not yet high enough for clinical use more, in typically developed children, intelligence and in diagnostics or treatment planning. However, exploring executive function correlate [Diamond, 2013]. the potential clinical utility of PGS is central in the pursuit Still, there are several unclear aspects related to cross- for precision medicine for NDs [Editorial, 2010; Torkamani, diagnostic features of NDs, which are behaviorally defined, Wineinger, & Topol, 2018]. PGS often only explain a few and their association to cognitive dysfunctions [Gillberg, percentiles of variation in disease status, meaning that they 2010]. Building on the recent progress in GWAS of different have limited predictive power. For many people with “aver- traits and disorders, the PGS has emerged as a tool that age risk,” quantifying the exact risk might be of limited enables investigation of the polygenic components of dif- value. However, identifying those with particularly low or ferent disorders and to explore the association between high risk may be of some value. A viable approach is to look genes, symptoms, and functioning. Furthermore, genetic at “extreme” PGS, as done by others [Andersson et al., 2018; underpinnings of general cognitive ability (intelligence) Khera et al., 2018; Schizophrenia Working Group of the Psy- can be used to identify differences in cognitive factors chiatric Genomics Consortium, 2014]. Individuals with between NDs [Savage et al., 2018], including executive high PGS for a given disorder might also be at increased risk function. This could provide a novel understanding of the for known comorbid traits and disorders. Thus, “extreme” underlying disease mechanisms, as outlined in the Research PGS analysis can be a promising tool to investigate the Domain Criteria initiative [Cuthbert, 2014]. Approaches to genetic contribution to symptom severity and disease char- dissect social and cognitive traits are also of clinical impor- acteristics in NDs. tance in the ASD field since children and adolescents are The key characteristics of ASD, deficits in social skills and admitted to specialist health care due to functional deficits, abnormal behaviors, strongly affect every day functioning mainly related to social and/or behavioral impairment. [de Vries & Geurts, 2015], which is often the primary rea- For people with ASD, structured situations with clear son for seeking professional help for families with ASD chil- expectations are easier than unstructured everyday situa- dren. Cognitive processes play an important role in the tions [Kenworthy, Yerys, Anthony, & Wallace, 2008]. Thus, development of social skills, and in moderating behavioral standardized and structured neuropsychological tests might responses [Gardiner & Iarocci, 2018], in particular executive not capture cognitive deficits important for everyday life function [Rommelse, Geurts, Franke, Buitelaar, & Hartman, functions [Kenworthy et al., 2008]. Actually, questionnaires 2011]. These abilities depend on response inhibition, inter- assessing executive functions have shown to have higher ference control, working memory, and flexibility [Fried- ecological validity than neuropsychological laboratory tests man & Miyake, 2017], which enable regulation of thought and might provide important information about how exec- and goal-directed behavior [Miyake et al., 2000]. Executive utive dysfunctions affect every day functioning. This find- dysfunction is suggested to be involved in the development ing applies to both clinical and nonclinical samples, as well of key symptoms and behaviors in ASD [Demetriou et al., as specifically to children and adults with ASD [Demetriou 2017; Hill, 2004] and other psychiatric disorders charac- et al., 2017; Tillmann et al., 2019; Vriezen & Pigott, terized by social dysfunctions and abnormal behavior, 2002]. The Behavior Rating Inventory of Executive Func- such as schizophrenia [Amann et al., 2012]. Importantly, tion (BRIEF) is one of the most used clinical tools to mea- social difficulties are associated with executive dysfunc- sure executive functions in everyday life [Gioia, Isquith, tions in everyday life in ASD [Leung, Vogan, Powell, Guy, & Kenworthy, 2015]. Furthermore, recent evidence Anagnostou, & Taylor, 2015; Torske, Naerland, Oie, suggests that the Social Responsiveness Scale (SRS) is a valid Stenberg, & Andreassen, 2017]. measure for social impairments in ASD [Constantino & ASD is a spectrum diagnosis with large variation in clini- Gruber, 2005]. Because the BRIEF and the SRS both are con- cal characteristics and functions. Children and adolescents tinuous scales, they can be used to study difficulties beyond with subthreshold symptoms who do not receive an ASD diagnostic categories [Geschwind & State, 2015]. Here, we

2 Torske et al./Polygenic scores and executive function INSAR used the BRIEF to measure executive functions and the SRS Table 1. Sample Characteristics (N = 176) to measure social skills as they occur in natural social N (%) settings. The primary aim of the present study was to determine Male/boys 134 (76.1%) Female/girls 42 (23.9%) if ASD PGS is associated with executive dysfunction in Mean (SD) Range everyday life in a sample of children and adolescents Age (years) 11.7 (3.7) 5–22 admitted for clinical assessment of ASD. Second, we aimed Full-scale IQ 93.7 (13.5) 70–133 to disentangle the polygenic components of NDs and cog- n nitive traits in a clinically relevant setting by investigating ASD 120 (67.8%) ASD with ADHD 42 how the PGS for ASD, ADHD, and general intelligence ADHD without ASD 26 (INT) are related to executive function. We hypothesized ASD diagnosis n = 120 n that high level of ASD PGS will be associated with worse Childhood autism (F84.0) 18 BRIEF scores beyond ASD diagnosis in a sample referred Atypical autism (F84.1) 5 for ASD assessment. Furthermore, since the participants in Asperger syndrome (F84.5) 56 Pervasive developmental 41 our sample have an ASD diagnosis and/or ASD symptom- disorder unspecified (F84.9) atology, we expected the ASD PGS to be more strongly associated with BRIEF scores than the PGS for ADHD and Low ASD PGS High ASD PGS P-value INT. We also investigated the association between the PGS Boys/girls n 18/8 n 22/5 0.573 and social skills in an everyday setting, and we hypothe- ASD n 16 n 20 0.618 sized that the PGS for ASD should be positively associated ADHD n 8 n 12 0.585 Full-scale IQ score (SD) 93.1 (14.6) 95.9 (10.7) 0.439 with social difficulties in everyday life.

ADHD, attention deficit/hyperactivity disorder; ASD, autism spectrum disor- der; IQ, intelligence quotient; SD,standarddeviation;PGS,polygenicscores; Method ASD polygenic groups Low and High based on ASD PGS at P-value <0.1. Participants

The total sample consisted of 176 children, adolescents, (F84.1), 56 with Asperger syndrome (F84.5) and 41 with and young adults who were referred to a specialized hos- Pervasive developmental disorder-not otherwise specified pital unit for NDs for evaluation of ASD. The current sam- (PDD-NOS) (F84.9). Of the 120 participants with ASD, ple was part of the national BUPGEN network and was 42 had a comorbid disorder of ADHD. In the subthresh- recruited between 2013 and 2018. For a detailed descrip- old ASD group, 26 had an ADHD diagnosis (See Table 1). tion of the recruitment procedure see Grove et al. supple- Of the participants without ASD or ADHD (n = 30), one mentary material [Grove et al., 2019]. Inclusion criteria had chronic motor or vocal tic disorder, four had Tourette were ASD-related difficulties and suspected ASD diagno- syndrome, two had mixed specific developmental disor- sis, independent of final ASD diagnosis. The participants ders, five had specific developmental disorders of speech were assessed by a team of experienced clinicians (clinical and language, two had specific developmental disorders psychologists and child psychiatrist). They assessed the of scholastic skills, two had other specified behavioral participants based on anamnestic information and des- and emotional disorders, two had disorder of psychologi- criptions of difficulties in everyday life activities using cal development, one had some other childhood disorder the gold standard tools, the Autism Diagnostic Observa- of social functioning and one had obsessive compulsive tion Schedule (ADOS), and/or the Autism Diagnostic disorder. Interview-Revised (ADI-R). Exclusion criteria were full In addition, we included 29 typically developed con- scale intelligence quotient (IQ) below 70. No participants trols (10 girls) with genetic and behavioral data available. had any significant sensory losses (vision and/or hearing) The controls were recruited from local schools through that could interfere with testing. The participants were in invitations/bulletins to all students/parents, and had no the age range 5–22 years and spoke Norwegian fluently. history of learning disabilities or psychiatric problems. Of the total sample of 176 participants, 120 fulfilled For a more detailed description of the controls see the diagnostic criteria for an ASD diagnosis, and the Hoyland et al. [2019]. remaining (n = 56) non-ASD diagnostic group was named We divided the clinical participants (across the ASD and the subthreshold ASD group. The subthreshold ASD group the subthreshold ASD groups) into three groups (Low, included participants with ASD symptomatology not Moderate, High) based on each of their PGS for ASD, reaching the criteria of ASD diagnosis, who may also have ADHD and INT. The Low, Moderate and High PGS groups other diagnoses. Of the 120 children and adolescents were identified based on the normal distribution curve of with ASD (91 boys and 29 girls), 18 were diagnosed with the standardized PGS scores: the Moderate group con- childhood autism (F84.0), five with atypical autism sisted of all participants with PGS within one standard

INSAR Torske et al./Polygenic scores and executive function 3 deviation from the mean, and the Low and High groups A t-score ≥76 represents severe interference in daily social respectively comprised participants with PGS in the lower interactions and is strongly associated with a clinical diag- tail and the upper tail of the distribution. This subdivision nosis of ASD [Constantino & Gruber, 2005]. Internal con- roughly corresponds to the participants with the 15% low- sistency is found to be high, both in population-based est scores constituting the Low group, those with the 15% samples and clinical samples (Cronbach’s α = 0.93–0.97) highest scores the High group, and everyone in between [Constantino & Gruber, 2005]. the Moderate group. In the linear regression, we used the groups Low, Moderate and High PGSs. The discovery sam- Behavior rating inventory of executive function. ples our PGS are based on consisted of 18,381 ASD cases The BRIEF is a 86 items questionnaire designed to assess exec- and 27,969 controls for the ASD PGS [Grove et al., 2019], utive function in everyday liferatedbyparents,teachers, 20,183 ADHD cases and 35,191 controls for the ADHD and/or self-report [Gioia, Isquith, Guy, & Kenworthy, PGS [Demontis et al., 2018] and 269,867 cases for the INT 2000]. In this study, we used the t-scores from the parents’ PGS [Savage et al., 2018]. report for the age group 5–18 years. For the few participants who were over the age of 18, we knew that they lived at Clinical Measures home with their parents and thattheirparentsthereforehad agoodbasisforassessingtheireverydayfunction.Thepar- The participants underwent a thorough clinical evalua- ents respond if the described behavior has been a problem tion for ASD using the gold standard tools, the ADOS for the child/adolescent the past 6 months by circling: and/or the ADI-R, administered by an experienced clini- (a) never, (b) sometimes, or (c) often. The raw scores were cian (psychiatrist and/or psychologist) [Lord et al., 2000; computed into the Software Portfolio (BRIEF SP) that has sep- Lord, Rutter, & Le Couteur, 1994]. In addition, a compre- arate norms based on respondent, age, and gender of the hensive diagnostic assessment was done based on ques- child/adolescent [Gioia et al., 2000]. The BRIEF consists of tionnaires, clinical interviews, naturalistic observations, three indexes and eight nonoverlapping subscales. The and formal testing of cognitive function. ADOS and/or Behavior Regulation Index (BRI) incorporates three subscales: ADI-R were performed in connection with clinical assess- inhibit, shift, and emotional control. The Metacognition ment for ASD. IQ was measured using the fourth version Index (MI) consists of five subscales: initiate, working mem- of Wechsler’s Intelligence Scale for Children [Wechsler, ory, plan/organize, organization of materials, and monitor. 2003] or another of Wechsler’s tests for the appropriate Together these eight subscales (three from BRI and five from age group [Wechsler, 2002, 2008]. The diagnostic assess- MI) form the Global Executive Composite Index (GEC). The ments of ADHD were done by a clinician (psychologist items that make up the BRI index are intended to measure and/or psychiatrist) specialized in child and adolescent the ability to modulate both behavioral and emotional con- psychology/psychiatry or a pediatrician, based on all pre- trol, and to move flexibly from one activity to another. The viously mentioned measures and according to formal MI index on the other hand is related to active problem solv- diagnostic criteria. ing, and the ability to initiate, organize, and monitor your own actions [Gioia et al., 2000]. A higher score is associated Social responsiveness scale. The SRS is a 65 items ques- with more problems related to executive function, and tionnaire made to identify the presence and severity of social t-scores ≥65 are considered to represent clinically significant impairment in natural social settings within the autism spec- levels. In our analysis, we used the continuous t-scores. The trum. A parent who is familiar with the individual’scurrent internal consistency is reported to be high (Cronbach’s behavior and developmental history responds to each ques- α =0.80–0.98) [Gioia et al., 2000]. tion on a Likert scale: (a) not true, (b) sometimes true, (c) often true, or (d) almost always true. The SRS consists of Polygenic Scores five treatment subscales; Social Awareness, Social Cognition, Social Communication, Social Motivation, and Autistic Man- Genotyping. The DNA was extracted with standard nerisms, which together form the total score [Constantino & methods either from blood samples or saliva collected in Gruber, 2005]. We used the parent rater scale for children/ the clinic. The genotypes for the study were obtained adolescents aged 4–18 years in our study [Constantino & with Human Omni Express-24 v.1.1 (Illumina Inc., San Gruber, 2005]. For the few participants who were over the age Diego, CA) at deCODE Genetics (Reykjavik, Iceland). of 18, we knew that they lived at home with their parents and Quality control was performed using PLINK 1.9 [Purcell that their parents therefore had a good basis for assessing their et al., 2007]. Briefly, variants were excluded if they had everyday function. We used continuous t-scores in our ana- low coverage (<95%), had low MAF (<0.01), deviated lyses and a higher score means more difficulties related to from Hardy–Weinberg equilibrium (P < 10−4), or occurred social function. A total score in the range 60–75 indicates clin- at significantly different frequencies in different ically significant deficits in social reciprocal interaction, and genotyping batches (FDR < 0.5). Whole individual geno- amildtomoderateinterferencein everyday interactions. types were excluded if they had low coverage (<95%) or

4 Torske et al./Polygenic scores and executive function INSAR high likelihood of contamination (heterozygosity above Statistical Analyses mean + 5 SD). MaCH software [Li, Willer, Ding, Scheet, & Abecasis, 2010] was used to obtain variant pseudo- Data analyses were conducted using the statistical package dosages (sums of imputation probabilities for the two IBM SPSS Statistics for Windows, version 25.0 (SPSS, Inc., haplotypes) for all participants based on reference haplo- Chicago, IL). Figure 1 was made using jamovi (Version 0.9). ’ types derived from the samples of European ancestry in We used Pearson sindependentt-tests to investigate the the 1,000 Genome Project. We had no challenges with mean differences in all clinical variables (MI, BRI, SRS, full- kinship in our sample (no relatedness above pi-hat 0.1 in scale IQ, and age) between the Low and the High PGS our sample). All our participants (controls and clinical) groups. Chi-square for crosstabs was used to investigate dif- were collected in the same time period, genotyped on the ferences in the distribution of sex and ASD and ADHD diag- same array, and imputed and processed together. Any noses in the Low and the High ASD PGS groups. We also ’ variants imputed with MAF lower than 0.05 or informa- used Pearson sindependentt-tests to investigate the differ- tion score lower than 0.8 were excluded from the subse- ences in PGS between controls and the clinical group to val- quent PGS calculations. idate the ASD PGS in our sample. Separate linear regression models were conducted to explain variance in BRIEF scores (BRI, MI, and GEC) across ASD, ADHD, and INT PGS Polygenic score. The variants’ effects were estimated from groups. Age, sex, SRS scores, full-scale IQ, and PCs were an inverse-variance weighted meta-analysis of the GWAS entered as covariates. We also used linear regression with summary statistics for all original sub-studies except any total SRS score as the dependent variable and ASD, ADHD sub-studies our own participants were drawn from. All sum- and INT PGS groups, age, sex, full-scale IQ, and BRIEF scores mary statistics were quality controlled by removing variants (BRI and MI) as independent variables. To describe the dis- that met any of the following conditions: minor allele fre- tribution of data, we added scatterplots of age by BRIEF quency <0.05; imputation quality INFO <0.8; not present scores and ASD PGS by BRIEF scores (See Supplementary in more than half of the sub-studies. The remaining vari- Figs. S1A-B and S2A-B) and Q-Q plots of BRI and MI scores ants were clumped into independent regions on the basis (See Supplementary Figs. S3A-B). The genetic principal com- of the linkage disequilibrium structure of the 1,000 ponents (PCs) that entered the linear regression model as Genomes Phase III European population. PLINK v1.9 was covariates of no interest were the ones with the highest cor- used with the following parameters: –clump-p1 1.0 –clump- relation with ASD, ADHD, and INT PGS (highest and p2 1.0 –clump-r2 0.2 –clump-kb 500. The allelic dosage unique correlation PGS ASD = PC02, PGS ADHD = PC01, coefficient (or logarithm of the odds ratio) of each of the and PGS INT = PC04) and the dependent variables variants with minimum P-values from all independent (BRIEF/GEC = PC06 and SRS = PC08, respectively). We used regions were taken as weights in constructing the PGSs. We standard procedures to control for population stratification, obtained GWAS summary statistics for ASD from Grove and we were unable to identify clear population subgroups et al. [2019], ADHD from Demontis et al. [2018] and INT from Savage et al. [2018] [Demontis et al., 2018; Grove et al., 2019; Savage et al., 2018]. The summary statistics were used to generate PGS, and test the association with spe- cificphenotypesinthecurrentsample, a procedure referred to as “genetic risk profiling” [Martin, Daly, Robinson, Hyman, & Neale, 2018]. We used a GWAS P-value threshold for inclusion of 0.1 for all PGS (ASD, ADHD, and INT). This P-value threshold was chosen because it resulted in the highest explained variance in ASD case/control status in the study our PGS was based on [Grove et al., 2019]. The scores included contributions from about 32,000 variants. The participants in our subsample and Grove et al. [2019], Demontis et al. [2018], and Savage et al. [2018] are all of European ancestry and are therefore genetically relatively homogeneous.

Copy-number variations. CNVs in molecular studies Figure 1. The ASD PGS scores for typically developed controls, fi have been identi ed as risk factors for ASD [Geschwind, the subthreshold ASD group, and the ASD group (ASD PGS at 2011]. Information about the CNVs in our sample was P < 0.1). The dots represent the mean scores, and the bands obtained from the chip genotypes, following standard inside the boxes are the median. ASD, autism spectrum disorder; CNV calling methodology [Sonderby et al., 2018]. PGS, polygenic score.

INSAR Torske et al./Polygenic scores and executive function 5 in our sample. For scatterplots of the PCs see Supplemen- Characteristics of the ASD PGS groups. In the total tary Figures S4A-D. The coordinates of the study partici- sample (n = 176), there were no significant group differ- pants in the relevant genetic PCs (PC1, 2, 3, 4, and 6) are ences between the Low and High ASD PGS groups on age inserted in the context of the European reference population (P = 0.059), full-scale IQ (P = 0.439), SRS total (P = 0.921), from the 1,000 genomes project in Supplementary Fig- or sex distribution (Pearson chi-square 1.113, P = 0.573). ures S5A-J. We also checked if the outliers had any substan- The Low and High ASD PGS groups did not, however, tial effect on the analysis and we tried to remove seven of reflect the diagnostic groups, and there was no signifi- the participants with the highest PC01 (outliers). The main cance difference between the distribution of ASD or findings were still the same (the difference in BRI score ADHD diagnosis (Pearson chi-square 0.962, P = 0.618 and between the high and low ASD groups was still significant) Pearson chi-square 0.107, P = 0.585) in the Low and High (P =0.036),andintheregressionanalysispredictingBRIthe PGS groups (See Table 1). We found no significant differ- same three covariates were significant; sex P =0.026,SRS ences in parental education level, which is highly related P =<0.001,andASDPGSgroupP =0.020;R2 =0.390).Fur- to income and socioeconomic status, between the Low thermore, we also fitted the regression models with continu- and the High ASD PGS groups (fathers education level ous PGS instead of PGS groups (See Supplementary P = 0.784, mothers education level P = 0.798). Tables S1A-C). Due to moderate sample size, we did not adjust alpha levels for multiple testing because of the risk of Validation of the PGS in the clinical group com- type 2 errors. All tests were considered significant with a pared to typically developed controls. To validate the two-sided P-value lower than 0.05. We report unadjusted ASD PGS, we compared the PGS from the ASD group P-values and are cautious in drawing our conclusions. (n =120),thesubthresholdASDgroup(n =56),andthetyp- ically developed controls (TDC) (n =29)usingPearson’s Ethical Considerations independent t-tests. The PGS for ASD differentiated between the TDC and the clinical group (ASD and subthreshold ASD The study was approved by the Regional Ethical Committee group, n =176;t(203) = 2.61, 95% CI [1.2 × 10−5, and the Norwegian Data Inspectorate (REK #2012/1967), 8.7 × 10−5], P =0.010.ThePGSforASDalsodifferentiated and was conducted in accordance with the Helsinki Declara- between the TDC and both the subthreshold ASD group; tion of the World Medical Association Assembly. Written t(83) = −2.07, 95% CI [−8.6 × 10−5, −1.7 × 10−6], P =0.042 informed consent was given from parents or legal guardians and the ASD group t(147) = −2.63, 95% CI [−9.1 × 10−5, for participants under 18 years.Participantsover18years −1.3 × 10−5], P =0.009.ThePGSforASDdidnotsignifi- gave written consent. cantly differ between those with ASD and subthreshold ASD group (P =0.578).Therewasnosignificant difference between the TDC and the clinical group (ASD and sub- Results threshold ASD group) on PGS for ADHD (P =0.205)orINT (P =0.944).TheTDCgrouphadsignificantly lower scores Clinical characteristics. In the total sample, there was a (less impairment/dysfunction) than the clinical group on significant difference between those who had an ASD diagno- both SRS and BRIEF (P <0.001),andwassignificantly older sis (n =120)andthesubthresholdASDgroup(n =56)on: than the clinical group (mean control group = 16.3 years, total SRS (P <0.001),totalt-score Global Executive Compos- mean clinical group = 11.7 years (ASD and subthreshold ite (GEC) from BRIEF (P =0.003),t-score BRI from the BRIEF ASD), P <0.001). Altogether the ASD PGS differentiated (P <0.001),andage(P =0.001).Therewerenosignificant dif- between the ASD group and TDC, and also between the sub- ferences between the groups on total IQ (P =0.841)ort-score threshold ASD group and the TDC (Fig. 1). MI from BRIEF (P =0.068).Therewerenosignificant differ- ences between those who had an ASD diagnosis and the sub- threshold ASD group on any of the PGS; ASD (P =0.578), PGS of ASD, ADHD, and INT and their association ADHD (P =0.810),andINT(P =0.842)(allP-bin 0.1). to executive function (BRIEF). There was a significant Those who had a comorbid diagnosis of ASD and ADHD group difference between the Low and the High ASD PGS (n =42)hadasignificantly higher score (were more groups in the BRI scores from the BRIEF; BRI t = 62.9 in impaired) on both the total t-score GEC (P =0.001)andthe the Low PGS group and t = 71.5 in the High PGS group; t-score MI (P =0.001)fromtheBRIEF,thanthosewhoonly t(49) = −2.51, 95% CI [−15.45, −1.71], P = 0.015, Cohen’s had an ASD diagnosis. There was no significant difference in d = 0.69. This means that the group with the highest PGS total SRS, total IQ, or BRI scores between those with an ASD for ASD had significantly more executive dysfunctions diagnosis only and those with comorbid ASD and ADHD. with BRI than the group with lowest PGS for ASD. We did In the whole sample (n = 176), there was a small, posi- not find a significant difference between the Low and tive correlation between ASD PGS and full-scale IQ, r =0 High ASD PGS groups in the MI scores from the BRIEF. We 0.033, but this was not significant (P = 0.673). also found no significant group differences between Low

6 Torske et al./Polygenic scores and executive function INSAR Table 2. Independent Sample t-Test: Differences in BRIEF Scores (BRI and MI) between the Low and the High PGS groups (ASD, ADHD, and INT) BRI Mean MIMean PGS group t-score(SD) n df t 95% CI P t-score(SD) n df t 95% CI P

Low ASD PGS 62.9 25 62.8 25 (12.6) (12.2) High ASD PGS 71.5 26 49 −2.51 [−15.45, −1.71] 0.015* 65.5 27 41.3a −0.94 [−8.57, 3.13] 0.354 (11.9) (8.1) Low ADHD 67.6 26 63.9 26 (14.9) (11.8) High ADHD PGS 64.9 26 50 0.74 [−4.64, 10.03] 0.465 67.0 26 50 −0.97 [−9.55, 3.32] 0.336 (11.2) (11.3) Low INT PGS 62.5 23 59.9 26 (17.1) (11.9) High INT PGS 66.2 26 47 −0.89 [−12.14, 4.71] 0.380 65.0 26 50 −1.61 [−11.31, 1.24] 0.113 (12.0) (10.6)

BRIEF-BRI, Behavior Rating Inventory of Executive Functions, Behavior Regulation Index; BRIEF-MI, Behavior Rating Inventory of Executive Functions, Metacognition Index; CI, confidence interval; df, degrees of freedom. Low and High ASD PGS: Autism spectrum disorder polygenic score subgroup (Low and High) at P < 0.1. Low and High ADHD PGS: Attention deficit/hyperactivity disorder polygenic score subgroup (Low and High) at P < 0.1. Low and High INT PGS: General intelligence polygenic score subgroup (Low and High) at P < 0.1. *P < 0.05 (two-tailed); **P < 0.01 (two-tailed). aEqual variance not assumed; Welch’s t-test. and High ADHD or INT PGS groups in either BRI or MI variance in the BRI score; F(11, 130) = 8.142, P < 0.001, scores from the BRIEF (See Tables 2 and 3). R2 = 0.41(unadjusted). Only SRS total (P = <0.001), PGS Furthermore, we investigated the presence of CNVs in ASD 0.1 group (P = 0.018), and sex (P = 0.022) made a sig- the Low and the High ASD PGS groups. Three of the par- nificant contribution to the model (See Table 4). This sug- ticipants in the Low ASD PGS group had CNVs, while no gests that the ASD PGS is more important than the CNVs were observed in the High ASD PGS group. ADHD PGS and the INT PGS in explaining executive dys- functions with behavior regulation (BRI) in our sample. Regression analyses with BRIEF as the dependent The same model explained 36% of the variance in MI variable. A linear regression model accounting for age, score; F(11, 134) = 6.836, P <0.001,R2 =0.36(unadjusted). sex, full-scale IQ, SRS total score, ASD, ADHD and INT In this case, only SRS total (P <0.001)madea significant PGS group as well as genetic PCs explained 41% of the contribution to the model. None of the PGS (ASD, ADHD,

Table 3. Independent Sample t-Test: Differences in BRIEF Scores (GEC) between the Low and the High PGS groups (ASD, ADHD, and INT) PGS group GECMean t-score (SD) n df t 95% CI P

Low ASD PGS 63.8 25 (12.7) High ASD PGS 69.0 26 49 −1.68 [−11.32, 1.00] 0.099 (9.0) Low ADHD PGS 66.5 26 (13.3) High ADHD PGS 67.7 26 50 −0.36 [−8.10, 5.64] 0.721 (11.2) Low INT PGS 61.9 25 (14.3) High INT PGS 66.2 26 40.4a −1.28 [−11.12, 2.49] 0.208 (9.1)

BRIEF-GEC, Behavior Rating Inventory of Executive Functions, Global Executive Composite; CI, confidence interval; df, degrees of freedom. Low and High ASD PGS: Autism spectrum disorder polygenic score subgroup (Low and High) at P < 0.1. Low and High ADHD PGS: Attention deficit/hyperactivity disorder polygenic score subgroup (Low and High) at P < 0.1. Low and High INT PGS: General intelligence polygenic score subgroup (Low and High) at P < 0.1. *P < 0.05 (two-tailed); **P < 0.01 (two-tailed). aEqual variance not assumed; Welch’s t-test.

INSAR Torske et al./Polygenic scores and executive function 7 Table 4. Linear Regression Model Summary: Prediction of Behavior Regulation Index (BRI) (Total sample n = 142a) 95% confidence interval

Predictor B SE B Lower Upper tPβ

Sex −5.21 2.25 −9.65 −0.77 −2.32 0.022* −0.17 Age −0.34 0.29 −0.92 0.23 −1.18 0.239 −0.08 Full-scale IQ 0.08 0.07 −0.06 0.22 1.13 0.261 0.08 SRS_total 0.52 0.06 0.39 0.65 8.10 <0.001** 0.59 ASD PGS subgr P < 0.1 4.07 1.70 0.71 7.44 2.40 0.018* 0.17 ADHD PGS subgr P < 0.1 −2.05 1.61 −5.23 1.13 −1.27 0.205 −0.09 INT PGS subgr P < 0.1 −0.61 1.79 −4.16 2.93 −0.34 0.732 −0.02

Model’s R2 = 0.408 (unadjusted). Models P-value = <0.001**. B, unstandardized regression coefficients; BRIEF-BRI, Behavior Rating Inventory of Executive Functions, Behavior Regulation Index; BRIEF-GEC, Behav- ior Rating Inventory of Executive Functions, Global Executive Composite; BRIEF-MI, Behavior Rating Inventory of Executive Functions, Metacognition Index; IQ, intelligence quotient; SE , standard error; SRS, Social Responsiveness Scale; β, standardized regression coefficients. ASD PGS subgroup P < 0.1: ASD polygenic groups low, moderate, and high based on the autism spectrum disorder polygenic score at P < 0.1. ADHD PGS subgroup P < 0.1: ADHD polygenic groups low, moderate, and high based on the attention deficit/hyperactivity disorder polygenic score at P < 0.1. INT PGS subgroup P < 0.1: INT polygenic groups low, moderate, and high based on the general intelligence polygenic score at P < 0.1. *P < 0.05 (two-tailed); **P < 0.01 (two-tailed). an corresponds to participants without any missing variables in outcomes or covariates (in the total sample of the study n = 176, there were 19 missing on SRS total and 11 missing on full-scale IQ). The covariates are included in the table to show their contribution to the model.

or INT) significantly contributed to this model (See Table 5). F(9, 132) = 9.753, P < 0.001, R2 = 0.40 (unadjusted). SRS These results indicate that the PGS did not have a significant total (P < 0.001), sex (P = 0.016), and ASD PGS (P = 0.021) role in explaining the metacognitive aspects of executive made a significant contribution to the model. function in our sample. For the regression model with GEC None of the PGSs (ASD, ADHD, or INT) had a significant as dependent variable see Table 6. contribution to the regression models, either with BRI or We also fitted regression models with each PGS sepa- MI as dependent variables, when we used the continuous rately to investigate their relationship to BRI and MI, PGS (BRI as dependent variable P = 0.234–0.488, MI as including sex, age, full-scale IQ, and total SRS as covariates. dependent variable P = 0.175–0.983) (See Tables S1A–C). The contribution of ASD PGS is mainly the same as when all three PGS were fitted in the same model. Fitting the The PGS association to social function (SRS). We regression model with BRI as dependent variable results in also investigated the association between the PGS and

Table 5. Linear Regression Model Summary: Prediction of Metacognition Index (MI) (Total sample n = 146a) 95% confidence interval

Predictor B SE B Lower Upper tPβ

Sex −2.45 1.87 −6.15 1.25 −1.31 0.192 −0.10 Age −0.17 0.23 −0.63 0.29 −0.72 0.471 −0.05 Full-scale IQ 0.05 0.06 −0.07 0.16 0.83 0.407 0.06 SRS_total 0.40 0.05 0.30 0.51 7.56 < 0.001** 0.57 ASD PGS subgr P < 0.1 1.36 1.41 −1.43 4.16 0.97 0.336 0.07 ADHD PGS subgr P < 0.1 0.99 1.34 −1.65 3.63 0.74 0.461 0.05 INT PGS subgr P < 0.1 1.15 1.42 −1.66 3.97 0.81 0.420 0.06

Model’s R2 = 0.359 (unadjusted). Models P-value < 0.001. ASD PGS subgroup P < 0.1: ASD polygenic groups low, moderate, and high based on the autism spectrum disorder polygenic score at P < 0.1. ADHD PGS subgroup P < 0.1: ADHD polygenic groups low, moderate, and high based on the attention deficit/hyperactivity disorder polygenic score at P < 0.1. INT PGS subgroup P < 0.1: INT polygenic groups low, moderate, and high based on the general intelligence polygenic score at P < 0.1. B, unstandardized regression coefficients; BRIEF-BRI, Behavior Rating Inventory of Executive Functions, Behavior Regulation Index; BRIEF-GEC, Behav- ior Rating Inventory of Executive Functions, Global Executive Composite; BRIEF-MI, Behavior Rating Inventory of Executive Functions, Metacognition Index; IQ, intelligence quotient; SE, standard error; SRS, Social Responsiveness Scale; β, standardized regression coefficients. *P < 0.05 (two-tailed) **P < 0.01 (two-tailed). an corresponds to participants without any missing variables in outcomes or covariates (in the total sample of the study n = 176, there were 19 missing on SRS total and 11 missing on full-scale IQ). The covariates are included in the table to show their contribution to the model.

8 Torske et al./Polygenic scores and executive function INSAR Table 6. Linear Regression Model Summary: Prediction of Global Executive Composite (GEC) (Total Sample n = 145a) 95% confidence interval

Predictor B SE B Lower Upper tPβ

Sex −3.64 1.86 −7.32 0.05 −1.95 0.053 −0.13 Age −0.32 0.23 −0.78 0.14 −1.38 0.170 −0.10 Full-scale_IQ 0.05 0.06 −0.07 0.17 0.86 0.389 0.06 SRS_total 0.48 0.05 0.37 0.59 8.90 < 0.001** 0.63 ASD PGS subgr P < 0.1 2.68 1.44 −0.18 5.53 1.85 0.066 0.13 ADHD PGS subgr P < 0.1 −0.02 1.35 −2.70 2.66 −0.01 0.989 0.00 INT PGS subgr P < 0.1 0.49 1.45 −2.37 3.35 0.34 0.736 0.02

R2 = 0.432 (unadjusted). Models P-value <0.001. B, unstandardized regression coefficients; BRIEF-BRI, Behavior Rating Inventory of Executive Functions, Behavior Regulation Index; BRIEF-GEC, Behav- ior Rating Inventory of Executive Functions, Global Executive Composite; BRIEF-MI, Behavior Rating Inventory of Executive Functions, Metacognition Index; IQ, intelligence quotient; SE, standard error; SRS, Social Responsiveness Scale; β, standardized regression coefficients. ASD PGS subgroup p0.1: ASD polygenic groups low, moderate, and high based on the autism spectrum disorder polygenic score at P < 0.1. ADHD PGS subgroup p0.1: ADHD polygenic groups low, moderate, and high based on the attention deficit/hyperactivity disorder polygenic score at P < 0.1. INT PGS subgroup p0.1: INT polygenic groups low, moderate, and high based on the general intelligence polygenic score at P < 0.1. *P < 0.05 (two-tailed); **P < 0.01 (two-tailed). an corresponds to participants without any missing variables in outcomes or covariates (in the total sample of the study n = 176, there were 19 missing on SRS total and 11 missing on full-scale IQ). The covariates are included in the table to show their contribution to the model. social function in everyday life measured with the SRS. highest and the lowest PGS for ASD. The participants in In an independent sample t-test we found a significant the High PGS group had on average a t-score of 71.5 on group difference between total SRS scores of the Low and BRI, which is in the clinical range of the scale. In compar- the High INT PGS groups; total t-score t =66.0intheLow ison, the participants in the Low PGS group had an aver- PGS group and t =76.2intheHighPGSgroup;t age t-score of 62.9, which is under the clinical cutoff (45) = −2.27, 95% CI [−19.11, −1.14], P =0.028,Cohen’s (Cohen’s d = 0.69). We did not observe any significant d =0.66.ThismeansthatthegroupwiththehighestPGS BRI difference between the Low and High PGS groups for for INT had significantly more problems related to social ADHD or INT. This finding was in line with our hypothe- function in everyday life than the group with Low PGS for sis that in a sample consisting of participants with an INT. We did not find any significant differences between ASD diagnosis and/or ASD symptomatology, the ASD the Low and High ASD PGS groups or the Low and High PGS would be more strongly associated with BRIEF scores ADHD PGS groups on the SRS total score (P =0.921and than the PGS for ADHD and INT. No significant differ- P =0.865).Furthermore,weperformedaregressionanalysis ences in the MI from the BRIEF were detected between and controlled for age, sex, full-scale IQ, MI from the BRIEF, the Low and High PGS groups for ASD, ADHD, or INT. BRI from the BRIEF, ASD, ADHD and INT PGS groups as Since ASD characteristics are quantitative traits, we well as PCs. The linear regression model was significant and included participants under the diagnostic threshold for explained 48% of the variance in total SRS score; F ASD. Furthermore, because of the large degree of comor- (12, 128) = 10.010, P <0.001, R2 =0.48(unadjusted).MI bidity in ASD, we included PGS of other diagnostic from the BRIEF (P = 0.001), BRI from the BRIEF (P < 0.001), groups. This enabled us to investigate the polygenic com- and sex (P = 0.019) had a significant contribution to the ponent of core ASD phenomena beyond diagnostic cate- model. None of the PGSs made a significant contribution gories. The use of extreme PGSs may already have clinical to the total SRS score (P = 0.101–0.742) in the regression relevance, for example, for cancer and cardiovascular dis- model. eases [Seibert et al., 2018; Torkamani et al., 2018]. In our study, we found a clinically meaningful and significant difference in the BRI scores from the BRIEF where the par- Discussion ticipants in the High ASD PGS had clinical/pathological t-scores and the Low ASD PGS group had nonclinical Our results showed a significant association between ASD score. Thus the use of extreme scores seems to have a PGS, representing the polygenetic components of ASD, potential for identifying clinically relevant levels of cog- and executive function deficits in everyday life in a clini- nitive problems. However, further studies are needed cal group seeking specialist health care due to autistic before clinical relevance can be established. symptomatology. In our study, the BRI from the BRIEF It is possible that polygenic factors in ASD could be differed significantly between individuals with the related to behavior differences, or other phenotypes not

INSAR Torske et al./Polygenic scores and executive function 9 related to executive function. Furthermore, executive children and adults, the mean correlation between scores on function deficits are not related to a specific disorder, but performance-based measures and behavioral ratings by use of characteristic of several NDs, with few deficits in typically BRIEF was reported to be 0.15 [Toplak, West, & Stanovich, developing controls. This can explain why PGS for ASD 2013]. Furthermore, it is known that genetic factors can con- in an ASD diagnostic group differs significantly from PGS tribute in complex ways to even performance-based tests for ASD in controls, but not significantly from PGS for intended to measure the same underlying construct like, for ASD in a non-ASD diagnostic group with ASD symptom- example, memory [Kremen et al., 2014]. In a recent meta- atology. Thus, in the current clinical sample referred for analysis of executive function in ASD, Demetriou et al. found autism symptomatology it is expected that the PGS for that most measures of executive function did not achieve ASD will be higher for the clinical group without ASD clinical utility in differentiating between ASD and typical than for controls. controls [Demetriou et al., 2017].However,informant-based The ASD PGS had a stronger relationship to executive measures based on BRIEF achieved absolute clinical marker function than the PGS for ADHD and INT in our clinical criteria. They conclude that the BRIEF is based on more repre- sample under assessment for ASD. This finding might sentative environmental situations and has therefore a higher imply that the common genetic variance associated with ecological validity than many performance-based measures, ASD is of greater importance for executive dysfunction and may thus be more appropriate in clinical practice. Fur- than the genetic variance associated with ADHD or INT. thermore, this is in line with recurrent findings that there is Even though the ADHD PGS and INT PGS were not sig- pronounced discrepancy between structured performance- nificant in the regression analysis with BRIEF as depen- based measures of general intelligence and adaptive function- dent variable, they still may have an association to BRIEF, ing (measures with e.g. Vineland-II) in ASD [Tillmann et al., and the findings must be interpreted with caution since 2019]. We therefore think that our finding that a higher ASD the regression analyses are based on null-hypothesis test- risk (higher ASD PGS) is associated with more executive func- ing. However, based on the actual t-scores in the Low and tion difficulties is novel and interesting, and is consistent the High groups we found that there is a significant dif- with the clinical representation of more executive function ference between the Low and the High ASD PGS groups difficulties in the ASD population. Our findings illustrate how in BRI, but not between Low and High ADHD and INT the polygenic component of NDs and its association to execu- PGS groups. Furthermore, our results indicate that the tive dysfunctions can disentangle psychological constructs behavioral regulation aspects of executive function are and may be used to explore possible underlying biological more strongly related to the polygenetic nature of ASD explanatory models of executive dysfunction. than the metacognitive aspects. The behavioral regula- In a GWAS study, Sun et al. found that ADHD children tion index from the BRIEF contains the subscales inhibi- had genetic variants related to behavioral regulation tion, flexibility, and emotional control. The subscale impairments measured with the BRIEF [Sun et al., flexibility is the hallmark of ASD, and it is within this 2018]. In children with NDs such as ASD and ADHD, area that those with ASD have the most pronounced diffi- there are indications that BRI from the BRIEF is more culties, also compared with other clinical groups with strongly associated with genetic factors than the meta- executive function deficits [Hovik et al., 2014]. Therefore, cognitive aspects of executive function [Sun et al., executive function deficits most specifically related to 2018]. This is in line with our findings of association ASD may also have the closest link to polygenetic compo- between the ASD PGS and the BRI subscale, but not the nents of ASD. metacognitive scale of BRIEF. Further, in our study we did Our finding that difficulties with the behavior regulatory not find any significant association between the PGS for aspects of executive function and ASD PGS are positively asso- ADHD and BRI. This seems to suggest a stronger genetic ciated contrasts Schork et al.’s finding that ASD PGS is associ- component of executive function in ASD in our sample, ated with better performance on a cognitive flexibility task as the effect size was bigger for ASD (d = 0.69) than for [Schork et al., 2018]. This might be explained by differences ADHD (d = 0.20). The lack of significant association in sample and method, as Schork et al. investigated typically between ADHD PGS and BRIEF scores may also be due to developing children with neuropsychological tests while we a smaller sample of ADHD participants, however the investigated children referredforclinicalassessmentand effect sizes for the extreme scores groups (ADHD PGS measured their executive functioning in everyday life. Fur- Low vs. High on BRI: d = 0.20 and MI: d = 0.27) support thermore, our finding is in line with studies of clinical the notion that ADHD PGS has less influence on BRIEF populations where executive function deficits are associated than the ASD PGS. Another explanation could be that with an ASD diagnosis [Demetriou et al., 2017]. The correla- ASD and ADHD involve different genetic mechanisms. tions between rating measures like the BRIEF and ASD and ADHD are both NDs and often manifest as performance-based measures of executive function are typi- comorbid conditions, both characterized by executive cally reported to be quite poor. In a review incorporating function deficits. In our study, the PGS for ASD groups did both clinical and nonclinical samples in addition to both not reproduce the diagnostic categories as both the Low

10 Torske et al./Polygenic scores and executive function INSAR and High PGS group consisted of children/adolescents with uncertain albeit in line with the “female protective ASD and/or ADHD. Dajani et al. [2016] studied a sample of model,” which proposes that more severe genetic muta- ASD, ADHD, or comorbid ASD and ADHD, and found that tions are required for a girl to develop ASD than for a boy executive function performance did not match the diag- [Levy et al., 2011]. Thus, common gene variants may also nostic categories. Sun et al. [2018] argue that is it more have different vulnerabilities depending on sex. important to evaluate executive function than diagnosis for targeting executive function interventions. Taken Strengths and limitations of the study. One of the together, this suggests that the Low and High PGS groups strengths of the current study is that it includes a clinically are not only an indirect measurement of ASD, but rather relevant sample of participants (n =176)whowerereferred reflect that polygenic disposition for ASD associated with for an ASD evaluation and were thoroughly clinically the behavior regulation part of executive function. Further- assessed for ASD. Furthermore, the PGSs are based on large more, it is important to be aware of potential factors that discovery samples with large power. Even though we have can influence the PGSs. In Grove et al.’s ASD sample it is relatively few participants, we found moderate effect sizes likely that some of the participants might have had comor- between measures of executive function (BRI) in the Low bid ADHD [Grove et al., 2019], and the ADHD risk score and High ASD PGS groups. The PGS for ASD is relatively calculated by Demontis et al. might be more specific to new and based on a smaller sample/cohort than the PGS for ADHD [Demontis et al., 2018]. However, it is likely that ADHD and INT. Yet, only the PGS for ASD is significantly there is some overlap between all the different phenotypes, associated with executive function deficits (BRI). However, especially at P-values as high as 0.1. because of a relatively small sample size, the uncorrected The INT PGS was significantly related to the amount of P-values, and the absence of an effect when using PGS as a social problems in this sample, and we found a moderate continuous variable, the results should be interpreted with effect size in the comparison of Low vs. High INT PGS on the caution. Furthermore, the findings are in need of replication SRS (P =0.028,d =0.66).ThePGSforASDorADHDdidnot in larger samples. We also had information about the pres- significantly influence the SRS score. This is in line with ence of possible CNVs in all the participants. We did not other studies linking common variants associated with high find more CNVs in the High vs. the Low ASD PGS group. IQ to social problems [Clarke et al., 2016; Grove et al., Therefore, it is not likely that CNVs are the reason for more 2019]. However, the contribution of INT PGS was not signifi- executive dysfunction in the High ASD PGS group. Another cant in our regression model. SRS measures autistic social strength of the study is that we had no challenges with kin- impairments in a quantitative score, and can also be viewed ship or ethnicity as confounding factors. as an impairment score. However, we did not find a signifi- Even though the participants were thoroughly assessed cant difference in total SRS score between the Low and High for ASD, we did not apply specific ADHD measures to com- ASD PGS groups. We might have too small and heteroge- pare the degree of ADHD symptoms; we only know neous sample consisting of different ASD diagnosis to detect whether they have a clinical diagnosis of ADHD. Although arelationshipbetweenASDPGSandSRS.Groveetal.found we controlled for age in the analyses, nonlinear age effects the Asperger diagnosis to be more strongly correlated to a could bias the results. high ASD PGS than classic autism (childhood autism) [Grove et al., 2019]. Clinical implications. It has been stated that “the end fi fi We did not nd a signi cant difference in full-scale IQ game of PGS […] is personalized medicine” [Zheutling & between the Low and the High ASD PGS groups in our sam- Ross, 2018]. Children and adolescent with high PGS for ple (P =0.439).Inthewholesample,therewasasmall,posi- ASD may be particularly vulnerable and have difficulties tive correlation between ASD PGS and full-scale IQ, with executive function. Despite the present small effect fi r =0.033,butthiswasnotsigni cant (P =0.673).Therefore, and predictive power, the findings suggest that PGS may fi our nding is not consistent with the reports showing high be a clinically useful tool in the future for children with risk for ASD associated with higher IQ in the general popula- NDs. If children at risk can be identified, it might be of tion [Clarke et al., 2016]. The reason for this is probably the clinical relevance to initiate prevention interventions size and heterogeneity of our sample, which groups together aimed at the executive difficulties, or stratify the more different ASD diagnoses. The Asperger diagnosis has previ- general ASD treatment by PGS. ously shown to be the ASD diagnosis with the highest posi- tive correlation with high IQ [Grove et al., 2019]. We found a significant contribution of sex to the Conclusions behavior regulation of executive function (BRI) in the regression model. This indicates that the relationship We report a significant relationship between PGS for ASD between genetic dispositions for ASD and executive and executive function deficits in terms of behavior regula- function may be different for boys and girls. The low tion in a clinical sample under evaluation for ASD symp- number of girls in our study makes these findings tomatology. Furthermore, we find that PGS of ASD, and not

INSAR Torske et al./Polygenic scores and executive function 11 ADHD nor INT, has a significant contribution to executive spectrum disorder. Journal of Autism and Developmental Disor- function deficits when controlling for confounders. To our ders, 45, 2734–2743. https://doi.org/10.1007/s10803-015-2438-1 knowledge, this is the first study to find an association Demetriou, E. A., Lampit, A., Quintana, D. S., Naismith, S. L., between the PGS for ASD and executive function in every- Song, Y. J. C., Pye, J. E., … Guastella, A. J. (2017). Autism day life, and shows how information from PGS can be used spectrum disorders: A meta-analysis of executive function. Molecular Psychiatry, 23, 1198–1204. https://doi.org/10.103 in a clinical neurodevelopmental sample. 8/mp.2017.75 Demontis, D., Walters, R. K., Martin, J., Mattheisen, M., Acknowledgments Als, T. D., Agerbo, E., … Neale, B. M. (2018). Discovery of the first genome-wide significant risk loci for attention deficit/- We are thankful to the BUPGEN partners and to all the par- hyperactivity disorder. Nature Genetics, 51, 63–75. https:// ticipants. The research was supported by the National doi.org/10.1038/s41588-018-0269-7 Research Council of Norway (Grant #213694), the South- Diamond, A. (2013). Executive functions. Annual Review of Psychol- Eastern Norway Regional Health Authority (Grant #39763), ogy, 64, 135–168. https://doi.org/10.1146/annurev-psych-113011- and Vestre Viken Hospital Trust (Grant #6903002). 143750 Friedman, N. P., & Miyake, A. (2017). Unity and diversity of executive functions: Individual differences as a window on References cognitive structure. Cortex, 86, 186–204. https://doi.org/10. 1016/j.cortex.2016.04.023 Amann, B., Gomar, J. J., Ortiz-Gil, J., McKenna, P., Sans- Gardiner, E., & Iarocci, G. (2018). Everyday executive function Sansa, B., Sarro, S., … Pomarol-Clotet, E. (2012). Executive predicts adaptive and internalizing behavior among children dysfunction and memory impairment in schizoaffective dis- with and without autism spectrum disorder. Autism Research, order: A comparison with bipolar disorder, schizophrenia and 11(2), 284–295. https://doi.org/10.1002/aur.1877 healthy controls. Psychological Medicine, 42(10), 2127–2135. Gaugler, T., Klei, L., Sanders, S. J., Bodea, C. A., Goldberg, A. P., https://doi.org/10.1017/S0033291712000104 Lee, A. B., … Buxbaum, J. D. (2014). Most genetic risk for American Psychiatric Association. (2013). Diagnostic and statisti- autism resides with common variation. Nature Genetics, cal manual of mental disorders (DSM-5). Washington, DC: 46(8), 881–885. https://doi.org/10.1038/ng.3039 American Psychiatric Association. Geschwind, D. H. (2011). Genetics of autism spectrum disorders. Andersson, E., Crowley, J. J., Lindefors, N., Ljotsson, B., Trends in Cognitive Sciences, 15(9), 409–416. https://doi.org/ Hedman-Lagerlof, E., Boberg, J., … Ruck, C. (2018). Genetics 10.1016/j.tics.2011.07.003 of response to cognitive behavior therapy in adults with Geschwind, D. H., & State, M. W. (2015). Gene hunting in major depression: A preliminary report. Molecular Psychiatry, autism spectrum disorder: On the path to precision medicine. 24, 484–490. https://doi.org/10.1038/s41380-018-0289-9 Lancet Neurology, 14(11), 1109–1120. https://doi.org/10. Benca, C. E., Derringer, J. L., Corley, R. P., Young, S. E., 1016/S1474-4422(15)00044-7 Keller, M. C., Hewitt, J. K., & Friedman, N. P. (2017). Gillberg, C. (2010). The ESSENCE in child psychiatry: Early Predicting cognitive executive functioning with polygenic symptomatic syndromes eliciting neurodevelopmental clini- risk scores for psychiatric disorders. Behavior Genetics, 47(1), cal examinations. Research in Developmental Disabilities, 11–24. https://doi.org/10.1007/s10519-016-9814-2 31(6), 1543–1551. https://doi.org/10.1016/j.ridd.2010.06.002 Clarke, T. K., Lupton, M. K., Fernandez-Pujals, A. M., Starr, J., Gioia, G. A., Isquith, P. K., Guy, S. C., & Kenworthy, L. (2000). Davies, G., Cox, S., … McIntosh, A. M. (2016). Common poly- Behavior rating rating inventory of executive function genic risk for autism spectrum disorder (ASD) is associated with (BRIEF). Florida: Psychological Assessment Resources, Inc. cognitive ability in the general population. Molecular Psychia- Gioia, G. A., Isquith, P. K., Guy, S. C., & Kenworthy, L. (2015). try, 21(3), 419–425. https://doi.org/10.1038/mp.2015.12 BRIEF 2: Behavior rating inventory of executive functions Constantino, J. N., & Gruber, C. P. (2005). Social responsiveness (2nd ed.). Florida: Psychological Assessment Resources, Inc. scale (SRS) manual. Los Angeles: Western Psychological Service. Grove, J., Ripke, S., Als, T. D., Mattheisen, M., Walters, R. K., Craig, F., Margari, F., Legrottaglie, A. R., Palumbi, R., de Won, H., … Borglum, A. D. (2019). Identification of com- Giambattista, C., & Margari, L. (2016). A review of executive func- mon genetic risk variants for autism spectrum disorder. tion deficits in autism spectrum disorder and attention-deficit/ Nature Genetics, 51(3), 431–444. https://doi.org/10.1038/ hyperactivity disorder. Neuropsychiatric Disease and Treatment, s41588-019-0344-8 12, 1191–1202. https://doi.org/10.2147/NDT.S104620 Hill, E. L. (2004). Executive dysfunction in autism. Trends in Cuthbert, B. N. (2014). The RDoC framework: Facilitating transi- Cognitive Sciences, 8(1), 26–32. tion from ICD/DSM to dimensional approaches that integrate Hovik, K. T., Egeland, J., Isquith, P. K., Gioia, G., Skogli, E. W., neuroscience and psychopathology. World Psychiatry, 13(1), Andersen, P. N., & Oie, M. (2014). Distinct patterns of every- 28–35. https://doi.org/10.1002/wps.20087 day executive function problems distinguish children with Dajani, D. R., Llabre, M. M., Nebel, M. B., Mostofsky, S. H., & tourette syndrome from children with ADHD or autism spec- Uddin, L. Q. (2016). Heterogeneity of executive functions trum disorders. Journal of Attention Disorders, 21, 811–823. among comorbid neurodevelopmental disorders. Scientific https://doi.org/10.1177/1087054714550336 Reports, 6, 36566. https://doi.org/10.1038/srep36566 Hoyland, A. L., Naerland, T., Engstrom, M., Torske, T., de Vries, M., & Geurts, H. (2015). Influence of autism traits and exec- Lydersen, S., & Andreassen, O. A. (2019). Atypical event-related utive functioning on quality of life in children with an autism potentials revealed during the passive parts of a Go-NoGo task

12 Torske et al./Polygenic scores and executive function INSAR in autism spectrum disorder: A case-control study. Molecular Psychology, 41(1), 49–100. https://doi.org/10.1006/cogp.1999. Autism, 10, 10. https://doi.org/10.1186/s13229-019-0259-3 0734 International Schizophrenia Consortium, Purcell, S. M., n.d. (2010). A decade for psychiatric disorders. Nature, 463 Wray, N. R., Stone, J. L., Visscher, P. M., O’Donovan, M. C., (7277), 9. https://doi.org/10.1038/463009a … Sklar, P. (2009). Common polygenic variation contributes Purcell, S., Neale, B., Todd-Brown, K., Thomas, L., Ferreira, M. A., to risk of schizophrenia and bipolar disorder. Nature, 460 Bender, D., … Sham, P. C. (2007). PLINK: A tool set for (7256), 748–752. https://doi.org/10.1038/nature08185 whole-genome association and population-based linkage ana- Kenworthy, L., Yerys, B. E., Anthony, L. G., & Wallace, G. L. (2008). lyses. American Journal of Human Genetics, 81(3), 559–575. Understanding executive controlinautismspectrumdisorders https://doi.org/10.1086/519795 in the lab and in the real world. Neuropsychology Review, 18(4), Ramaswami, G., & Geschwind, D. H. (2018). Genetics of autism 320–338. https://doi.org/10.1007/s11065-008-9077-7 spectrum disorder. Handbook of Clinical Neurology, 147, Khera, A. V., Chaffin, M., Aragam, K. G., Haas, M. E., Roselli, C., 321–329. https://doi.org/10.1016/B978-0-444-63233-3.00021-X Choi, S. H., … Kathiresan, S. (2018). Genome-wide polygenic Rommelse, N. N., Geurts, H. M., Franke, B., Buitelaar, J. K., & scores for common diseases identify individuals with risk Hartman, C. A. (2011). A review on cognitive and brain endo- equivalent to monogenic mutations. Nature Genetics, 50(9), phenotypes that may be common in autism spectrum disor- 1219–1224. https://doi.org/10.1038/s41588-018-0183-z der and attention-deficit/hyperactivity disorder and facilitate Kremen, W. S., Panizzon, M. S., Franz, C. E., Spoon, K. M., the search for pleiotropic genes. Neuroscience and Biobehav- Vuoksimaa, E., Jacobson, K. C., … Lyons, M. J. (2014). ioral Reviews, 35(6), 1363–1396. https://doi.org/10.1016/j. Genetic complexity of episodic memory: A twin approach to neubiorev.2011.02.015 studies of aging. Psychology and Aging, 29(2), 404–417. Savage, J. E., Jansen, P. R., Stringer, S., Watanabe, K., Bryois, J., https://doi.org/10.1037/a0035962 de Leeuw, C. A., … Posthuma, D. (2018). Genome-wide asso- Lai, M. C., Lombardo, M. V., & Baron-Cohen, S. (2014). Autism. ciation meta-analysis in 269,867 individuals identifies new Lancet, 383(9920), 896–910. https://doi.org/10.1016/s0140- genetic and functional links to intelligence. Nature Genetics, 6736(13)61539-1 50(7), 912–919. https://doi.org/10.1038/s41588-018-0152-6 Leung, R. C., Vogan, V. M., Powell, T. L., Anagnostou, E., & Schizophrenia Working Group of the Psychiatric Genomics Con- Taylor, M. J. (2015). The role of executive functions in social sortium. (2014). Biological insights from 108 schizophrenia- impairment in Autism Spectrum Disorder. Child Neuropsychol- associated genetic loci. Nature, 511(7510), 421–427. https:// ogy, 22, 1–9. https://doi.org/10.1080/09297049.2015.1005066 doi.org/10.1038/nature13595 Levy, D., Ronemus, M., Yamrom, B., Lee, Y. H., Leotta, A., Schork, A. J., Brown, T. T., Hagler, D. J., Thompson, W. K., Kendall, J., … Wigler, M. (2011). Rare de novo and transmit- Chen, C. H., Dale, A. M., … Genetics, S. (2018). Polygenic risk ted copy-number variation in autistic spectrum disorders. for psychiatric disorders correlates with executive function in Neuron, 70(5), 886–897. https://doi.org/10.1016/j.neuron.20 typical development. Genes, Brain and Behavior, 18, e12480. 11.05.015 https://doi.org/10.1111/gbb.12480 Li, Y., Willer, C. J., Ding, J., Scheet, P., & Abecasis, G. R. (2010). Seibert, T. M., Fan, C. C., Wang, Y., Zuber, V., Karunamuni, R., MaCH: Using sequence and genotype data to estimate haplo- Parsons, J. K., … Practical Consortium. (2018). Polygenic haz- types and unobserved genotypes. Genetic Epidemiology, ard score to guide screening for aggressive prostate cancer: 34(8), 816–834. https://doi.org/10.1002/gepi.20533 development and validation in large scale cohorts. BMJ, 360, Lord, C., Elsabbagh, M., Baird, G., & Veenstra-Vanderweele, J. j5757. https://doi.org/10.1136/bmj.j5757 (2018). Autism spectrum disorder. Lancet, 392(10146), Snyder, H. R., Miyake, A., & Hankin, B. L. (2015). Advancing 508–520. https://doi.org/10.1016/S0140-6736(18)31129-2 understanding of executive function impairments and psy- Lord, C., Risi, S., Lambrecht, L., Cook, E. H., Jr., Leventhal, B. L., chopathology: Bridging the gap between clinical and cogni- DiLavore, P. C., … Rutter, M. (2000). The autism diagnostic tive approaches. Frontiers in Psychology, 6, 328. https://doi. observation schedule-generic: A standard measure of social org/10.3389/fpsyg.2015.00328 and communication deficits associated with the spectrum of Sonderby, I. E., Gustafsson, O., Doan, N. T., Hibar, D. P., Martin- autism. Journal of Autism and Developmental Disorders, 30 Brevet, S., Abdellaoui, A., … 16p11.2 European Consortium, (3), 205–223. for the ENIGMA-CNV working group. (2018). Dose response Lord, C., Rutter, M., & Le Couteur, A. (1994). Autism Diagnostic of the 16p11.2 distal copy number variant on intracranial Interview-Revised: A revised version of a diagnostic interview volume and basal ganglia. Molecular Psychiatry, 1–19. for caregivers of individuals with possible pervasive develop- https://doi.org/10.1038/s41380-018-0118-1 mental disorders. Journal of Autism and Developmental Dis- Sun, X., Wu, Z., Cao, Q., Qian, Y., Liu, Y., Yang, B., … Wang, Y. orders, 24(5), 659–685. (2018). Genetic variant for behavioral regulation factor of Martin, A. R., Daly, M. J., Robinson, E. B., Hyman, S. E., & executive function and its possible brain mechanism in atten- Neale, B. M. (2018). Predicting polygenic risk of psychiatric tion deficit hyperactivity disorder. Scientific Reports, 8(1), disorders. Biological Psychiatry, 86, 97–109. https://doi.org/ 7620. https://doi.org/10.1038/s41598-018-26042-y 10.1016/j.biopsych.2018.12.015 Tillmann, J., San Jose Caceres, A., Chatham, C. H., Crawley, D., Miyake, A., Friedman, N. P., Emerson, M. J., Witzki, A. H., Holt, R., Oakley, B., … EU-AIMS LEAP Group. (2019). Investi- Howerter, A., & Wager, T. D. (2000). The unity and diversity gating the factors underlying adaptive functioning in autism of executive functions and their contributions to complex in the EU-AIMS Longitudinal European Autism Project. Autism “Frontal Lobe” tasks: A latent variable analysis. Cognitive Research, 12(4), 645–657. https://doi.org/10.1002/aur.2081

INSAR Torske et al./Polygenic scores and executive function 13 Toplak, M. E., West, R. F., & Stanovich, K. E. (2013). Practitioner Supplementary Figure S1A. Scatterplot age by BRI review: do performance-based measures and ratings of execu- t-score from the BRIEF. tive function assess the same construct? Journal of Child Psy- Supplementary Figure S1B. Scatterplot age by MI chology and Psychiatry, 54(2), 131–143. https://doi.org/10. t-score from the BRIEF. 1111/jcpp.12001 Supplementary Figure S2A. Scatterplot BRI t-score Torkamani, A., Wineinger, N. E., & Topol, E. J. (2018). The personal from the BRIEF by ASD PGS group P < 0.1. and clinical utility of polygenic risk scores. Nature Reviews Genet- Supplementary Figure S2B. Scatterplot MI t-score ics, 19(9), 581–590. https://doi.org/10.1038/s41576-018-0018-x Torske, T., Naerland, T., Oie, M. G., Stenberg, N., & from the BRIEF by ASD PGS group P < 0.1. Andreassen, O. A. (2017). Metacognitive aspects of executive Supplementary Figure S3A. Q-Q plots for BRI t-scores function are highly associated with social functioning on from the BRIEF. parent-rated measures in children with autism spectrum dis- Shapiro–Wilk test P =0.089.Kurtosis=−0.469. Skew- order. Frontiers in Behavioral Neuroscience, 11, 258. https:// ness = −0.136. doi.org/10.3389/fnbeh.2017.00258 Supplementary Figure S3B. Q-Q plots for MI t-scores Vriezen, E. R., & Pigott, S. E. (2002). The relationship between from the BRIEF. parental report on the BRIEF and performance-based mea- Shapiro–Wilk test P =0.208.Kurtosis=−0.246. Skew- sures of executive function in children with moderate to ness = −0.227. severe traumatic brain injury. Child Neuropsychology, 8(4), Supplementary Figures S4A–D. Scatterplots principal 296–303. https://doi.org/10.1076/chin.8.4.296.13505 components. Wechsler, D. (2002). Wechsler preschool and primary scale of intelligence - Third edition. San Antonio, TX: Pearson. (PC01 by PC02, PC01 by PC03, PC01 by PC04 and PC01 Wechsler, D. (2003). Wechler Intelligence Scale for Children - by PC06). Fourth edition. San Antonio, TX: Pearson. Supplementary Figures S5A–J. Study participants rel- Wechsler, D. (2008). Wechsler Adult Intelligence Scale–Fourth evant PCs in the context of the European reference Edition. San Antonio, TX: Pearson. population. Zheutling, A. B., & Ross, D. (2018). Polygenic risk scores: What Supplementary Table S1A. Linear regression model are they good for? Biological Psychiatry, 83(11), e51–e53. summary. Prediction of Behavior Regulation Index (BRI). https://doi.org/10.1016/j.biopsych.2018.04.007 Total sample n = 1421. Supplementary Table S1B. Linear regression model Supporting Information summary. Prediction of Metacognition Index (MI). Total sample n = 1461. Additional supporting information may be found online Supplementary Table S1C. Linear regression model in the Supporting Information section at the end of the summary. Prediction of Global Executive Composite article. (GEC). Total sample n = 1451.

14 Torske et al./Polygenic scores and executive function INSAR Supplementary Figure S1A. Scatterplot age by BRI t-score from the BRIEF

Supplementary Figure S1B. Scatterplot age by MI t-score from the BRIEF

Supplementary Figure S2A. Scatterplot BRI t-score from the BRIEF by ASD PGS group p<0.1

Supplementary Figure S2B. Scatterplot MI t-score from the BRIEF by ASD PGS group p<0.1

Supplementary Figure S3A. Q-Q plots for BRI t-scores from the BRIEF

Shapiro-Wilk test p = 0.089. Kurtosis = -0.469. Skewness = -0.136.

Supplementary Figure S3B. Q-Q plots for MI t-scores from the BRIEF

Shapiro-Wilk test p = 0.208. Kurtosis = -0.246. Skewness = -0.227.

Supplementary Figures S4A-D. Scatterplots principal components (PC01 by PC02, PC01 by PC03, PC01 by PC04 and PC01 by PC06)

Supplementary Figures S5 A-J. Study participants relevant PCs in the context of the European reference population