physical activity and sedentary behaviour in association with
LIKES – Research Reports on Sport and Health 292 PHYSICAL ACTIVITY AND SEDENTARY BEHAVIOUR IN ASSOCIATION WITH ACADEMIC PERFORMANCE AND COGNITIVE FUNCTIONS IN SCHOOL-AGED CHILDREN HEIDI SYVÄOJA Academic dissertation to be publicly discussed, by permission of the Faculty of Social Sciences of the University of Jyväskylä, in the Seminarium S212, on November 15th, 2014, at 12 noon LIKES – Research Center for Sport and Health Sciences Jyväskylä 2014 University of Jyväskylä Faculty of Social Sciences Department of Psychology Author’s address Heidi Syväoja LIKES – Research Center for Sport and Health Sciences Viitaniementie 15a, FI-40720 Jyväskylä, Finland heidi.syvaoja@likes.fi Supervisors Professor Timo Ahonen Department of Psychology University of Jyväskylä Research Director Tuija Tammelin LIKES – Research Center for Sport and Health Sciences Doctor Marko Kantomaa LIKES – Research Center for Sport and Health Sciences Imperial College London, UK Reviewers Professor Charles H. Hillman Department of Kinesiology and Community Health University of Illinois, Urbana-Champaign, Illinois Professor Urho Kujala Department of Health Sciences University of Jyväskylä Opponent Professor Charles H. Hillman Department of Kinesiology and Community Health University of Illinois, Urbana-Champaign, Illinois LIKES – Research Reports on Sport and Health 292 ISBN | ISSN (print) 978-951-790-367-7 | 0357-2498 ISBN | ISSN (pdf) 987-951-790-368-4 | 2342-4788 Editor Tuija Tammelin Distribution LIKES – Research Center for Sport and Health Sciences Viitaniementie 15a, 40720 Jyväskylä, Finland Copyright © 2014, Heidi Syväoja & LIKES – Research Center for Sport and Health Sciences Printing Grano, Jyväskylä ABSTRACT Syväoja, Heidi Physical activity and sedentary behaviour in association with academic performance and cognitive functions in school-aged children LIKES – Research Reports on Sport and Health 292 Jyväskylä: LIKES – Research Center for Sport and Health Sciences, 2014. In addition to physical health benefits, physical activity may enhance children’s cognitive and academic performance. Excessive sedentary behaviour, in turn, may have harmful effects on cognitive functions and academic achievement in children and adolescents. The purpose of this study was to determine the associations of physical activity and sedentary behaviour with academic achievement and cognitive functions in school-aged children. In addition, this study aimed to evaluate the internal consistency and one-year stability of the seven tests of computerized neuropsychological test battery used to assess children’s cognitive functions. Two hundred seventy-seven children from five schools in the Jyväskylä school district in Finland (58% of the 475 eligible students; mean age 12.2 years; 56% girls) participated in the study in the spring of 2011. Children’s physical activity and sedentary behaviour were self-reported and measured objectively using accelerometers. Academic achievement scores (teacher-rated grade point averages) were provided by the education services of the city of Jyväskylä. Cognitive functions were evaluated with Cambridge Neuropsychological Test Automated Battery (CANTAB) using two tests for visual memory, three test for executive functions and two tests for attention. During spring 2012, the follow-up measurements were conducted among 74 children. Self-reported physical activity had an inverse curvilinear association with academic achievement, and total screen time had a linear negative association with academic achievement, whereas objectively measured physical activity or sedentary time had no association with academic achievement. High levels of objectively measured physical activity were associated with good performance in attentional reaction time test. High levels of objectively measured sedentary time was associated with good performance in sustained attention test. Objectively measured physical activity or sedentary time had no association with other domains of cognitive functioning. Self-reported physical activity, total screen time or TV viewing had no association with assessed cognitive functions. High self-reported time spent in video game play and computer use was associated with poor performance in the tests measuring visuospatial working memory and shifting and flexibility of attention, respectively. The one-year stability of most cognitive tests was moderate-to-good, but the internal consistency was below an acceptable level for most of the tests, highlighting the need to confirm the psychometric characteristics of the computerized tests among target populations. The results of this study showed that physical activity was positively – and screen time negatively – associated with academic achievement and certain cognitive functions in children. In addition, the positive association with objectively measured sedentary time and sustained attention shows that sedentary time also includes activities that may benefit certain cognitive functions. The results highlight the importance of promoting a physically active lifestyle. Keywords: Physical activity, sedentary behaviour, academic achievement, cognitive functions, internal consistency, stability, children. TIIVISTELMÄ Syväoja, Heidi Liikunnan ja liikkumattomuuden yhteydet lasten kognitiiviseen toimintaan ja koulumenestykseen Liikunnan ja kansanterveyden julkaisuja 292 Jyväskylä: LIKES-tutkimuskeskus, 2014. Liikunta saattaa terveyshyötyjensä lisäksi vaikuttaa myönteisesti myös lasten kognitiiviseen ja akateemiseen suoriutumiseen. Liiallinen ruutuaika voi puolestaan heikentää lasten ja nuorten kognitiivista toimintaa ja koulumenestystä. Tämän tutkimuksen tarkoituksena oli selvittää liikunnan ja liikkumattomuuden yhteyksiä koulumenestykseen ja kognitiiviseen toimintaan kouluikäisillä lapsilla. Lisäksi selvitettiin lasten kognitiivisten toimintojen mittaamiseen käytetyn neuropsykologisen testipatteriston seitsemän eri testin luotettavuutta ja pysyvyyttä. Tutkimukseen osallistui 277 lasta viidestä eri Jyväskylän alueen koulusta (58 prosenttia kutsutuista; keski-ikä 12,2 vuotta; 56 prosenttia tyttöjä) keväällä 2011. Lapset täyttivät liikuntaa ja ruutuaikaa koskevan kyselyn. Liikunnan ja liikkumattoman ajan määrää mitattiin myös objektiivisesti kiihtyvyysanturilla. Lasten koulumenestystiedot (todistuksen arvosanat) kerättiin Jyväskylän kaupungin opintorekisteristä. Kognitiivisia toimintoja mitattiin tietokonepohjaisella CANTAB-testipatteristolla (Cambridge Neuropsychological Test Automated Battery), johon valittiin kaksi testiä mittaamaan muistia, kolme testiä mittaamaan toiminnanohjausta ja kaksi testiä mittaamaan tarkkaavaisuutta. Osa lapsista (n=74) osallistui seurantamittauksiin keväällä 2012. Itseraportoitu liikunta oli myönteisesti ja ruutuaika käänteisesti yhteydessä koulumenestykseen, kun taas objektiivisesti mitattu liikunta ja liikkumaton aika eivät olleet yhteydessä koulumenestykseen. Runsas objektiivisesti mitattu liikunta oli yhteydessä parempaan reaktioaikaan tarkkaavaisuustestissä. Runsas objektiivisesti mitattu liikkumaton aika oli yhteydessä parempaan pitkäkestoiseen tarkkaavaisuuteen. Objektiivisesti mitattu liikunta ja liikkumaton aika eivät olleet yhteydessä muihin mitattuihin kognitiivisen toiminnan osa-alueisiin. Runsas itseraportoitu videopelien peluu oli yhteydessä heikompaan suoriutumiseen työmuistitehtävässä ja tietokoneen käyttö heikompaan suoriutumiseen tarkkaavaisuuden joustavuutta mittaavassa testissä. Itseraportoitu liikunta, kokonaisruutuaika tai television katselu eivät olleet yhteydessä mitattuihin kognitiivisen toiminnan osa-alueisiin. Kognitiivisten testien mittaustulosten pysyvyys vuoden välein mitattuna vaihteli kohtalaisesta hyvään, kun taas eri testien sisäinen konsistenssi oli suhteellisen heikko suurimmassa osassa testejä, minkä vuoksi tietokonepohjaisten testipatteristojen psykometriset ominaisuudet tulisi tarkistaa aina kohdejoukossa. Tulosten mukaan liikunta on myönteisesti ja ruutuaika käänteisesti yhteydessä koulumenestykseen ja tiettyihin kognitiivisen toiminnan osa-alueisiin. Kuitenkaan kaikki liikkumattomuus ei ole samanarvoista kognition kannalta, vaan osa liikkumattomasta ajasta saattaa sisältää toimintoja, jotka ovat kognition kannalta hyödyllisiä. Tutkimus antaa tukea liikunnallisen elämäntavan edistämiseen koulumenestyksen ja kognitiivisen toiminnan näkökulmasta. Avainsanat: Liikunta, liikkumattomuus, koulumenestys, kognitiivinen toiminta, luotettavuus, pysyvyys, lapset. No sensible decision can be made any longer without taking into account not only the world as it is but also the world as it will be. -Isaac Asimov ACKNOWLEDGEMENTS This study was carried out at the LIKES – Research Center for Sport and Health Sciences, Jyväskylä, and at the Department of Psychology, University of Jyväskylä between 2011 and 2014. I would like to express my sincere thanks and appreciation to all those who have contributed to or otherwise participated in this work. I wish to express my deepest gratitude to my three indispensable supervisors: Professor Timo Ahonen, PhD, from the Department of Psychology, University of Jyväskylä, for his wisdom, encouragement, and warm support throughout the whole project; Research Director Tuija Tammelin, PhD, from the LIKES Research Center, for her unending enthusiasm and the way she trusted me and introduced me to the scientific world; Doctor Marko Kantomaa, PhD, from the LIKES Research Center and Department of Epidemiology and Biostatistics, MRC–HPA Centre for Environment and Health, Imperial College London, UK, for his positive attitude, guidance and help whenever I needed it. I have been privileged to work with this valuable team. I sincerely thank Professor Charles H. Hillman, PhD, from Department of Kinesiology and Community Health, University of Illinois, Urbana-Champaign, Illinois, and Professor Urho Kujala, PhD, from the Department of Health Sciences, University of Jyväskylä, the official reviewers of this doctoral thesis, for their constructive criticism and valuable comments that have improved the quality of the thesis. I also wish to thank other co-authors for their valuable contributions to this work: Professor Asko Tolvanen, PhD, from the Department of Psychology, University of Jyväskylä, for his full professionalism and significant guidance in the area of statistical approaches. Sincere thanks to Pekka Räsänen, Lic. in psychology, clinical neuropsychologist, Niilo Mäki Institute, for valuable comments and important discussions about the neuropsychological testing. I owe my warmest gratitude to Anna Kankaanpää and Harto Hakonen, from the LIKES Research Center, for their patient assistance and priceless help in the field of statistics and data analysis. I want to express warm thanks to Virpi Inkinen, Kirsti Siekkinen and Janne Kulmala, from the LIKES Research Center, for their contribution and commitment to data collection. Many thanks also go to Annaleena Aira, Martta Walker, and Maija Mörsky, from the LIKES Research Center, for their help in the area of informing. I wish to thank the director Eino Havas and the staff of the LIKES – Research Center for Sport and Health Sciences, Jyväskylä, for creating a unique working environment. Special thanks go to my colleagues, Katariina Kämppi and Jaana Kari, for sharing the thoughts and moods every day in the office. I am very grateful to all my friends and relatives, especially to Hanna-Kaisa, Eveliina and girls in our dance team, for giving me something else to think about and for sharing fun and not always so fun moments during these years. Finally, I express my feeling of thankfulness and appreciation to my dear parents, Ritva and Antti, for their unconditional love and unstinting support throughout every step of my life. I wish to thank my dear sisters, Henna and Henni, with all my heart, for sharing important moments in my life, and for helping me to overcome setbacks. I owe my love and deepest gratitude to Janne for giving me his unfailing love, sympathy, and support. I love you all! Jyväskylä 1.11. 2014 Heidi Syväoja This study was financially supported by the Finnish Ministry of Education and Culture and the Research Programme on the Future of Learning, Knowledge and Skills (TULOS), Academy of Finland (grant 273971). LIST OF ORIGINAL PUBLICATIONS The thesis is based on the following original publications, which are referred to in the text by their Roman numerals. I Syväoja, H.J., Kantomaa, M.T., Ahonen, T., Hakonen, H., Kankaanpää, A. & Tammelin, T.H. 2013. Physical activity, sedentary behavior, and academic performance in Finnish children. Medicine and Science in Sports and Exercise 45 (11), 2098–2104. II Syväoja, H.J., Tammelin, T.H., Ahonen, T., Räsänen, P., Tolvanen, A., Kankaanpää, A. & Kantomaa, M.T. Internal consistency and stability of the CANTAB neuropsychological test battery in children. Submitted manuscript. III Syväoja, H.J., Tammelin, T.H., Ahonen, T., Kankaanpää, A. & Kantomaa, M.T. 2014. The Associations of Objectively Measured Physical Activity and Sedentary Time with Cognitive Functions in School-aged Children. PloS one 9 (7): e103559. ABBREVIATIONS ADHD B BDNF Beta CANTAB CFI CI CRF FIML GPA HBSC HRR ICC IED IRT MET MLR MPA MVPA OR p PA PACER PE PRM RCT RMSEA RTI RVP SD SE SOC SRM SRMR SSP TLI UK US VPA WHO ∆R2 attention deficit hyperactivity disorder unstandardized coefficient (in Table 5), estimate (in Table 6) brain-derived neurotrophic factor standardized coefficient Cambridge Neuropsychological Test Automated Battery comparative fit index confidence interval cardiorespiratory fitness full information maximum likelihood grade point average Health Behaviour in School-aged Children heart rate reserve intra-class correlations Intra-Extra Dimensional Set Shift item response theory metabolic equivalent robust standard errors moderate physical activity moderate to vigorous physical activity odds ratio p-value physical activity Progressive Aerobic Cardiovascular Endurance Run physical education Pattern Recognition Memory randomized controlled trial root mean square error of approximation Reaction Time Rapid Visual Information Processing standard deviation (in Tables 1–3) standard error (in Tables 5–6) Stockings of Cambridge Spatial Recognition Memory standardized root-mean-square residual Spatial Span Tucker–Lewis Index United Kingdom United States vigorous physical activity World Health Organization change in R square CONTENTS ABSTRACT TIIVISTELMÄ ACKNOWLEDGEMENTS LIST OF ORIGINAL PUBLICATIONS ABBREVIATIONS CONTENTS 1 INTRODUCTION ......................................................................................................................13 1.1 Physical activity and sedentary behaviour among school-aged children ....................................................................................................... 14 1.2 Academic achievement and cognitive functions ..................................... 16 1.2.1 Academic achievement..................................................................... 16 1.2.2 Cognitive functions ........................................................................... 18 1.2.3 Assessments of cognitive functions ................................................ 19 1.3 Associations of physical activity, academic achievement and cognitive functions ..................................................................................................... 20 1.3.1 Physical activity in association with academic achievement........ 21 1.3.2 Physical activity in association with cognitive functions .............. 24 1.4 Associations of sedentary behaviour, academic performance and cognitive functions ..................................................................................... 27 1.4.1 Sedentary behaviour in association with academic achievement 28 1.4.2 Sedentary behaviour in association with cognitive functions ...... 30 1.5 Summary of literature ............................................................................... 31 2 AIMS OF THE STUDY .............................................................................................................33 3 MATERIAL AND METHODS ................................................................................................35 3.1 Study population and data collection ....................................................... 35 3.2 Measurements ............................................................................................ 39 3.2.1 Self-reported physical activity and screen time............................. 39 3.2.2 Objective measures of physical activity and sedentary time ........ 39 3.2.3 Academic performance .................................................................... 40 3.2.4 Cognitive functions ........................................................................... 40 3.2.4.1 Visual memory ................................................................................. 41 3.2.4.2 Executive function ......................................................................... 41 3.2.4.3 Attention ............................................................................. 41 3.2.5 Potential confounders ...................................................................... 42 3.3 Ethics statement ......................................................................................... 42 3.4 Analytical strategies................................................................................... 42 3.4.1 Study I ................................................................................................ 42 3.4.2 Study II............................................................................................... 43 3.4.3 Study III ............................................................................................. 44 4 AN OVERVIEW AND THE MAIN RESULTS OF THE ORIGINAL STUDIES...........45 4.1 Study I: Physical activity, sedentary behaviour, and academic performance in Finnish children .............................................................. 45 4.2 Study II: Internal consistency and stability of the CANTAB neuropsychological test battery in children ............................................ 48 4.3 Study III: The associations of objectively measured physical activity and sedentary time with cognitive functions in school-aged children ......... 52 5 DISCUSSION ..............................................................................................................................55 5.1 Physical activity, academic achievement and cognitive functions ........ 55 5.1.1 Possible mechanisms explaining the associations between physical activity, academic achievement and cognitive functions ............................................................................................ 57 5.1.2 Differences between objectively measured and self-reported physical activity in terms of their association with academic achievement and cognitive functions ............................................. 57 5.2 Sedentary behaviour, academic achievement and cognitive functions. 58 5.3 Methodological considerations ................................................................. 60 5.3.1 General strengths and limitations ................................................... 60 5.3.2 Measurement of physical activity and sedentary behaviour ........ 61 5.3.3 Measurement of academic achievement ........................................ 61 5.3.4 Measurement of cognitive functions – Reliability and stability of cognitive test battery (CANTAB) ..................................................... 62 6 CONCLUSIONS..........................................................................................................................65 6.1 Summary and main conclusions ............................................................... 65 6.2 Implications and future directions ........................................................... 66 REFERENCES…………………………………………………………………………………………………...69 APPENDICES.…………………………………………………………………………………...………………89 ORIGINAL PUBLICATIONS ..……………………………………………………………………………113 1 INTRODUCTION Along with numerous societal changes over the past few decades, our lifestyle has become increasingly sedentary. This phenomenon is not only found among adults, but also in children (Nelson et al. 2006), who on average spend 4–8 hours per day being sedentary (Pate et al. 2011). In addition, only one-third of children are sufficiently active according to current physical activity recommendations (Ekelund, Tomkinson & Armstrong 2011). In Finland, about one half of primary school pupils and only one sixth of lower secondary school pupils fulfill the minimum recommendation for physical activity (Tammelin, Laine & Turpeinen 2013). Lack of physical activity is seen as one of the increasing risk factors for lifestyle diseases: physical inactivity is associated with higher levels of obesity, metabolic and cardiovascular risk factors, depression symptoms, and lower physical fitness in children, whereas adequate physical activity may benefit all of these risk factors (Mountjoy et al. 2011, Tremblay et al. 2011b, Ekelund et al. 2012). Children’s physical fitness has decreased and obesity has increased globally during the past couple decades (Tomkinson & Olds 2007, Lakshman, Elks & Ong 2012), including in Finland (Kautiainen et al. 2002, Huotari et al. 2010). Along with convincing research evidence on the significance of regular physical activity for health, the idea that physical activity may enhance and support learning has also emerged in recent years. There is evidence that physical activity enhances cognitive functions, especially executive functions in the elderly (Hillman, Erickson & Kramer 2008, Guiney & Machado 2013). In addition, there is a recent increase in research suggesting that physical activity benefits children’s cognitive functions and academic performance (Donnelly et al. 2009, Davis et al. 2011, Fisher et al. 2011, Kamijo et al. 2011, Chaddock-Heyman et al. 2013, Ardoy et al. 2014). In turn, excessive sedentary behaviour, especially screen-based sedentary behaviour, has been linked to poorer academic performance (Sharif & Sargent 2006, Mößle et al. 2010) and elevated risk of attention and learning difficulties (Swing et al. 2010, Weis & Cerankosky 2010). Physical activity may have an underestimated and underused potential to support learning. However, evidence of the favourable effects of physical activity and the harmful effects of sedentary behaviour on cognitive functions and academic achievement in healthy children and adolescents is still inconsistent and based on scarce research data. Moreover, most of the previous studies have used selfreported measures of physical activity and sedentary behaviour, with only a few using objectively measured physical activity or sedentary time in connection with educational outcomes and cognitive functions. As the rates of childhood physical inactivity are increasing worldwide, systematic research on the effects of physical activity and sedentary life on cognitive functions and academic performance is needed for both practitioners and policy makers to promote learning, education and health. This study aims to determine how subjectively and objectively measured physical activity and sedentary behaviour are associated with academic performance and cognitive functions in elementary school-aged children. 13 1.1 Physical activity and sedentary behaviour among schoolaged children According to traditional definition, physical activity is any bodily movement, which is produced by the contraction of skeletal muscle and increases energy expenditure above a basal level (Caspersen, Powell & Christenson 1985). Physical activity can be classified according to intensity using metabolic equivalent (MET) as a reference. One MET refers to the rate of energy expenditure while sitting at rest (US Department of Health and Human Services & US Department of Health and Human Services 2008). Therefore, moderate to vigorous physical activity (MVPA) is activity that increases the rest energy expenditure threefold (≥3 METs) (World Health Organization 2010), while small-to-large increases in breathing and heart-rate are caused by such activities as brisk walking, bicycling, running and swimming (US Department of Health and Human Services & US Department of Health and Human Services 2008). In addition to energy expenditure, physical activity can be defined as biocultural behaviour: it occurs in many forms and contexts that are strongly influenced by culture (Malina 2001). Exercise is a subcategory of physical activity. It is planned, structured and repetitive bodily movement that aims to improve or maintain one or more components of physical fitness (Caspersen, Powell & Christenson 1985). Sedentary behaviour can be defined as any waking behaviour requiring low levels of energy expenditure (≤ 1.5 METs) while in a sitting or reclining posture (Barnes et al. 2012). Sedentary behaviour includes many different activities, such as watching television, playing video games, computer use, reading, desk-based work, doing homework, and sitting while socializing. Sedentary behaviour and physical activity are independent constructs, as a person can have a great amount of sedentary time, but still meet the physical activity guidelines (Pate et al. 2011, Pate, O'Neill & Lobelo 2008). Physical activity guidelines for health benefits state that children (aged 5–11 years) and youth (aged 12–17 years) should accumulate at least 60 minutes of MVPA per day, including vigorous-intensity activities at least three days per week and activities that strengthen muscle and bone at least three days per week (World Health Organization 2010, Tremblay et al. 2011c). In addition, recent American guidelines for physical activity during school day, state that children should have opportunity to accomplish more than half of the recommended 60 minutes per day of MVPA during regular school hours provided by school district, administrators, teachers and parents (Kohl & Cook 2013). Furthermore, Canadian guidelines on sedentary behaviour recommend that children and youth should minimize the time spent being sedentary each day and especially limit recreational screen time to no more than two hours per day (Tremblay et al. 2011a). Finnish physical activity recommendations for school-aged children are consonant with these international guidelines (Tammelin & Karvinen 2008). As indicated before, the majority of children and youth both internationally and nationally do not meet these recommendations (Ekelund, Tomkinson & Armstrong 2011, Salmon et al. 2011). Exacerbating these issues, physical activity continues to decrease and sedentary time continues to increase from childhood to adolescence (Troiano et al. 2008). 14 Traditionally, physical activity and sedentary behaviour have been measured subjectively with self-reported questionnaires or diaries. For children and youth, physical activity questionnaires typically assess specific types of physical activity, but also recreational physical activity, mostly expressed as duration of physical activity, time spent in MVPA or estimates of energy expenditure (Ekelund, Tomkinson & Armstrong 2011). Questionnaires have been accepted as being capable of ranking inter-individual differences in physical activity and demonstrating correlations between physical activity and other variables (Strath et al. 2013, Shephard 2014). However, one of the main weaknesses of self-reports is accuracy, especially among children (Ekelund, Tomkinson & Armstrong 2011). Children may overestimate their physical activity levels, compared to objectively measured physical activity (Corder et al. 2010). Questionnaires used to evaluate sedentary behaviour have usually assessed the time spent in screen-based sedentary behaviour, especially time spent watching television (Tremblay et al. 2011b). In addition, non-screen-based sedentary behaviour such as educational sedentary behaviour, sedentary hobbies and social sedentary behaviours have been measured in self-reports (Pate et al. 2011). Recollection is also a potential issue for self-reports of sedentary behaviour. However, according to Pate et al. (2011), the estimated times spent in sedentary behaviour are similar between self-reports and objectively measured studies. Objective measurements of physical activity and sedentary time have improved the ability to accurately determine the volume, intensity, duration and frequency of children’s physical activity (Ekelund, Tomkinson & Armstrong 2011, Strath et al. 2013), eliminating bias caused by issues of recollection and social desirability and overcoming challenges caused by language and literacy difficulties (Evenson et al. 2006). Accelerometers are the most commonly used tools for objectively measuring physical activity and sedentary time (Ekelund, Tomkinson & Armstrong 2011). Accelerometers record acceleration and deceleration of the body. Raw accelerometer data is often converted into other units, such as activity counts (Strath et al. 2013). Activity counts can be used to determine the intensity of activity by using thresholds, which have been developed by calibrating activity counts with measurements of oxygen consumption. (Evenson et al. 2006). There is ongoing debate about the thresholds for different intensity categories (sedentary, light, moderate or vigorous) and how the lack of consensus influences interpretation of data (Ekelund, Tomkinson & Armstrong 2011, Strath et al. 2013). In addition, accelerometers cannot track all activities, such as cycling or actions that require lifting a load (Strath et al. 2013); this is one weakness of accelerometers. Usually, both questionnaires and objective measurements assess the total amount (frequency, duration and intensity) of physical activity and sedentary behaviour, not content or context. In this study, both self-report and accelerometer measurements were used to assess the amount of children’s physical activity and sedentary behaviour when determining the relationship of physical activity and sedentary behaviour with academic performance and cognitive functions. 15 1.2 Academic achievement and cognitive functions 1.2.1 Academic achievement Learning may be seen as a life-sustaining and an inevitable part of human growth and development. It is an active and interactive process, which results in changes in behaviour, but also in the skills, knowledge and emotional reactions underlying behaviour. Learning occurs in a specific cultural and social context. Many theories have been advanced over the years to understand how people learn. The social theory of learning was chosen to form the basis of this thesis. According to the social theory of learning, learning is seen as social participation (Bandura & McClelland 1977). This participation shapes who we are, what we do and how we interpret what we do (Figure 1) (Wenger 1998, 3–11, Wenger 2000). Learning as doing Learning as belonging Community Practice Learning Identity Meaning Learning as experience FIGURE 1. Learning as becoming Components of the social theory of learning. (Figure modified after Wenger 1998, 5). At schools, learning is the centre of action. Students’ learning is evaluated, monitored and supported. Academic achievement is one way of assessing how well children have achieved their educational goals. The purpose of assessment is to support and enhance learning and teaching. Internationally, different standardized tests – including the disciplines of reading, spelling, arithmetic and science – are used to measure academic achievement, and grades given by teachers at the end of the term are also used. (Rasmussen & Laumann 2013). In Finland, national standardized tests are not in use in primary or secondary schools. Assessment is based on teacher-rated academic achievement scores, which describe the level of performance in relation to the objectives outlined by the National Core Curriculum for Basic Education (National core curriculum for basic education 2004). Objectives not only include skills and knowledge, but also behaviour and students’ working skills, such as the ability to plan, carry out and evaluate their 16 own work. Children’s working abilities make up part of the subject-based assessment, but behaviour is assessed separately. Teachers’ evaluate children’s academic achievements scores independently. (Ouakrim-Soivio 2013). There is ongoing debate about the best possible method for grading, which is both comparable and commensurable. Grading that strongly depends on standardized achievement tests enables uniform evaluation and comparison between students. Standardized tests are practical, when measuring large group of children quickly, efficiently and affordably. (Ouakrim-Soivio 2013). However, it has been criticized for directing students to rehearse response techniques and for excluding incorrect responses rather than supporting actual knowledge and skills (Koretz & Hamilton 2006, Hamilton, Stecher & Yuan 2012). In addition, standardized tests have been criticized for being unable to measure children’s learning comprehensively enough, for example they are not suitable for estimating children’s ability to produce own stories (Ansley 1997). Grading, which is administered by the teacher and based on objectives outlined beforehand, enables a more versatile evaluation of children’s learning. However, if the objectives of the national curriculum for each grade are not carefully followed in the process of evaluation, a teacher’s independent role as evaluator may result in discrepancies in grades between students from different classes or schools with the same level of competence. Inequity and bias in grading weakens the opportunity to directly compare children’s academic achievements. (Ouakrim-Soivio 2013). From a psychological perspective, there are a lot of factors that influence and are associated with academic achievement (Winne & Nesbit 2010). The classic metaanalytic study by Wang, Haertel and Walberg (1994) showed that factors that directly influence children and children’s close social interactions had a greater impact on academic achievement than indirect influences. Children’s metacognitive and cognitive skills, classroom management, home environment and parental support, and social interactions between students and teachers had the greatest effect on academic achievement (Wang, Haertel & Walberg 1994). Cognitive processes, especially executive functions (defined later) (Lu et al. 2011, McClelland & Cameron 2011, Weber et al. 2013), have been shown to be strong predictors of academic achievement in elementary school. Motivation, and ultimately, achievement are influenced by people’s beliefs about their capability to do different tasks, but also the reasons for doing them (Wigfield & Cambria 2010). When individuals have high self-efficacy beliefs (strong beliefs about their own capabilities) (Caprara et al. 2011, Weber et al. 2013), they are more likely to engage in activities, persist in spite of difficulties, and succeed (Wigfield & Cambria 2010). Reasons for performing tasks – such as achievement values, goal orientations and interest, and intrinsic value – are also important predictors of achievement outcomes and learning (Wigfield & Cambria 2010, Weber et al. 2013). In addition, parental involvement in their children’s education (Karbach et al. 2013) also plays an important role in predicting children’s academic achievement. This study focused on teacher-rated academic scores, on the basis of which grade point averages (GPA) were calculated in order to represent children’s overall academic performance. 17 1.2.2 Cognitive functions From the neurocognitive perspective, learning is the reconstruction of neuronal networks. These neuronal networks form the biological foundation of cognition, intelligence and learning (Anderson 1997). Intelligence can be defined as an individual’s comprehensive cognitive ability to reason, plan, solve problems, think abstractly, understand complex ideas, learn quickly and learn from experience (Neisser et al. 1996, Nisbett et al. 2012). In addition, several specific concepts of cognition have been defined such as executive functions, memory and attention. Executive functions (also called executive control, the central executive, or cognitive control) are the collection of higher-order cognitive processes that control goal-directed actions. Essential for purposeful behaviour, executive functions regulate cognition, emotion and action. Three core executive functions are inhibition, working memory and mental flexibility. (Diamond 2013). Inhibitory control (inhibition) includes selective attention and the inhibition of inappropriate or interfering responses. In other words, inhibition enables appropriate action by controlling attention, behaviour, thoughts, and emotions to suppress dominant, automatic, or prepotent responses. (Best, Miller & Jones 2009, Diamond 2013). Working memory is defined as the ability to actively retain information in one’s mind and mentally manipulate it for brief periods of time. Working memory can be considered as a threecomponent system involving a control system of limited attentional capacity (a part of executive functions) and two assisting temporary storage systems (Baddeley & Hitch 1974, Baddeley 2012, Diamond 2013). Mental flexibility (shifting), in turn, is defined as the ability to change perspectives and shift between mental states, operations or tasks. These core functions are tightly linked together, and higher executive functions like reasoning, problem-solving and planning are built from these core functions. (Best, Miller & Jones 2009, Diamond 2013). According to one definition, memory is a lasting representation reflected in thought, experience, or behaviour. It is not a single unitary system, but consists of subsystems called working memory, short-term and long-term memory, which can also be divided into different components. Short-term memory refers to simple temporary storage of information, whereas working memory points to active manipulation of shortly maintained information (Baddeley 2012). As stated above, working memory consists of a supervisory system, which uses two short-term memory systems: one concerning speech and sound (phonological loop) and the other visuospatial information (visuospatial sketchpad) (Baddeley & Hitch 1974, Baddeley 2012, Diamond 2013). Long-term memory includes stored memories and learned knowledge and skills. It can be divided into explicit memory, which includes perceptual, episodic and semantic memories, and implicit memory, which includes implicit knowledge, primed biases and goals as well as highly practised habits and motor skills (Jeneson & Squire 2011, Baddeley 2012). Attention can be divided into three sub-systems: alerting, orienting and the executive network. Alerting involves achieving and maintaining optimal vigilance during a task, while orienting concerns the ability to prioritize sensory input by selecting a modality or location. The executive network, in turn, is active when processing targets, conflicts and errors (Petersen & Posner 2012, Posner & Rothbart 2014). Attention can be voluntary, goal-directed executive attention or spontaneous bottom-up attention. Executive attention (also called selective or focused attention) 18 enables us to selectively attend to and focus on what we decide, as well as to suppress attention to other stimuli. Spontaneous attention, in turn, is driven by the properties of these stimuli. Due to the selective and controlling nature of executive attention, it can be seen as one executive function, as described above. (Diamond 2013). Even though many cognitive tests evaluate one specific cognitive function, cognitive functioning is usually a dynamic interactive process of many different functions. For example, paying attention to sensory or internal information is directed by selective attention, after which information becomes available for working memory. Working memory can actively manipulate information, while inhibitory control hinders internal and external distractions. Managing information in working memory is determined by our goals and our existing information (in long-term memory), and attending to selective aspects of the environment. Manipulated information is encoded into long-term memory, and learning takes place. (Baars & Gage 2010, 33–60). From a developmental perspective, these neurocognitive functions emerge during the early years of life. However, especially the executive functions continue to develop throughout childhood and adolescence, due to protracted development of the brain structures that support them. Different domains of executive functions vary in their developmental trajectories and may not be fully functionally mature by the age of 12 years. (Luciana & Nelson 1998, Luciana & Nelson 2002, Best & Miller 2010). The focus of this thesis is to study associations of physical activity and sedentary behaviour on visual memory, executive functions and attention. Therefore, other domains of cognitive functions will not be discussed further. 1.2.3 Assessments of cognitive functions Traditionally, cognitive functions have been measured with paper-and-pencil tests to assess specific cognitive ability, for example, the Corsi Blocks test for measuring visuospatial working memory (Milner 1971), Tower of London measuring planning and spatial working memory (Shallice & Shallice 1982, Owen et al. 1990) and the Wisconsin Card Sorting test for measuring mental flexibility (Heaton et al. 1993). In recent years, the use of computer technology in neuropsychological assessments has increased worldwide, and computer-based test batteries have been used for psychological assessments, for evaluating changes over time, and for evaluating the effects of interventions (Lowe & Rabbitt 1998). Computer technology has made valuable contributions to neuropsychological assessments by increasing efficiency, ease and standardization of administration; reducing errors during scoring; and increasing accuracy of timing and response latencies (Cernich et al. 2007, Parsey & Schmitter-Edgecombe 2013). In this study, the Cambridge Neuropsychological Test Automated Battery (CANTAB) was used to assess children’s cognitive functions. The CANTAB is one of the oldest computer-based test batteries used to evaluate neurocognitive functions, particularly in clinical trials research. CANTAB tests are mainly based on traditional neuropsychological tests. These non-verbal tests measure visual and spatial memory, working memory, planning, different aspects of attention, and other areas of cognition. (Cambridge Cognition Ltd., 2006). According to previous reports, CANTAB has been found suitable for assessing cognitive functions in 4- to 90-year-old 19 individuals (Lowe & Rabbitt 1998, Luciana 2003). In addition, it has been found sensitive to cognitive deficits due to several neuropsychological and psychiatric conditions and diseases, especially in the elderly (e.g. Sahakian et al. 1993, Robbins et al. 1998, Blackwell et al. 2004, Levaux et al. 2007) but also in children (Luciana et al. 1999, Gau & Shang 2010, Rhodes et al. 2011, Fried et al. 2012). However, earlier studies did not provide adequate information about the reliability of CANTAB tests (Luciana & Nelson 2002). Luciana (2003) reported that internal consistency coefficients for CANTAB tests were high (0.73 – 0.95) in 4–12year-old children. Furthermore, studies measuring the test-retest reliability of CANTAB have been sparse. Lowe and Rabbit (1998) reported that in an elderly adult population, the test-retest agreement for the CANTAB tests was either moderate, ranging from 0.70 to 0.86, or low, ranging from 0.09 to 0.68. According to Fisher et al. (2011), intra-class correlations for CANTAB subtests, Spatial Span length, and Working Memory errors were quite low (ICC = 0.51 – 0.59) in healthy children. However, Gau and Shang (2010) reported higher intra-class correlations for CANTAB tests (Intra-Extra Dimensional Set Shift, Spatial Span, Spatial Working Memory, Stockings of Cambridge), ranging from 0.55 to 0.94 in a group of 10 children with attention deficit hyperactivity disorder (ADHD). To our knowledge, there are no other studies that establish the internal consistency agreement or test-retest agreement for CANTAB tests in a child population (Luciana 2003, Henry & Bettenay 2010), while the results of existing studies are, to some extent, inconsistent. 1.3 Associations of physical activity, academic achievement and cognitive functions Previous studies have shown that physical activity enhances neurocognitive functions and protects against neurodegenerative diseases in the elderly (Kramer & Erickson 2007, Hillman, Erickson & Kramer 2008, Lautenschlager et al. 2008, Erickson et al. 2011). Most typically the effects have been seen in executive functioning, related to selective attention and inhibitory control, mental flexibility and working memory (Guiney & Machado 2013). More recently, physical activity has been linked to better academic performance (Shephard 1997, Trost 2008, Trudeau & Shephard 2008, Trudeau & Shephard 2010, Centers for Disease Control and Prevention. 2010, Singh et al. 2012) and enhanced cognitive functions (Sibley & Etnier 2003, Hillman, Erickson & Kramer 2008, Hillman, Kamijo & Scudder 2011, Guiney & Machado 2013) in children. Earlier studies have reported that physical activity may benefit (Tuckman & Hinkle 1986, Tremblay, Inman & Willms 2000, Dwyer et al. 2001, Coe et al. 2004, Coe et al. 2006, Fredericks, Kokot & Krog 2006, Hillman et al. 2006, Nelson et al. 2006, Davis et al. 2007, Tremarche, Robinson & Graham 2007) or do not compromise (Low 1990, Sallis et al. 1999, Dollman, Boshoff & Dodd 2006, Yu et al. 2006, Ahamed et al. 2007, Sigfusdottir, Kristjansson & Allegrante 2007) academic and cognitive performance, but conflicting results have also been observed (Themane et al. 2006). In addition, acute exercise during the school day has been shown to have 20 benefits on academic and cognitive performance (McNaughten & Gabbard 1993, Catering & Polak 1999, Maeda & Randall 2003). Moreover, physical fitness (Sollerhed & Ejlertsson 1999, Kim et al. 2003, Grissom 2005, Hillman, Castelli & Buck 2005, Castelli et al. 2007) and participation in sports (Silliker & Quirk 1997, Dexter 1999, Lindner 1999, Stephens & Schaben 2002, Eitle 2005) has been linked to enhanced academic achievement and cognition, but not in all studies (Schumaker, Small & Wood 1986, Hanson & Kraus 1998, Daley & Ryan 2000, Eitle & Eitle 2002, Lindner 2002, Miller et al. 2005). In 2010, the results of these earlier studies were synthesized and published by the Centers for Disease Control and Prevention. A comprehensive report summarizes the scientific literature (43 articles) published between 1985 and October 2008 that concerns the associations of school-based physical activity and academic performance (Centers for Disease Control and Prevention 2010). School-based physical activity included physical education, physical activity during recess, classroom physical activity and extracurricular activity, while academic performance included academic achievement, academic behaviour and indicators of cognitive skills and attitudes. According to the report, physical activity had a positive association (50.5% of the associations summarized), no association (48% of the associations summarized), or negative association (1.5% of the associations summarized) with academic performance. Moreover, increased physical activity on school days was not associated with attenuated academic performance (Centers for Disease Control and Prevention 2010). The associations of physical activity, academic achievement and cognitive functions observed in studies published in 2008 or later are described in more detail below. 1.3.1 Physical activity in association with academic achievement A few intervention studies have reported that integrated physical activity in academic lessons (Donnelly et al. 2009, Reed et al. 2010), increased physical education (Ericsson 2008, Spitzer & Hollmann 2013, Ardoy et al. 2014) and aerobic exercise programmes (Davis et al. 2011) benefit children’s academic achievement. In the study of Donelly et al. (2009), academic achievement scores for reading, spelling and math significantly improved from the baseline to three years in children participating in the intervention, compared to control children. In addition, improvements in total academic achievement have been reported (Ardoy et al. 2014), as well as in math (Ericsson 2008, Davis et al. 2011, Spitzer & Hollmann 2013, Ardoy et al. 2014), mother tongue (Ericsson 2008, Spitzer & Hollmann 2013) and social studies (Reed et al. 2010). However, in some of these studies, physical activity had no effect on achievements of reading and language (Reed et al. 2010, Davis et al. 2011, Ardoy et al. 2014), foreign language (Spitzer & Hollmann 2013) or science (Reed et al. 2010). To summarize the results of these few intervention studies, increasing physical activity to the school week benefits academic achievement in certain school subjects in certain studies, but has no effect on certain subjects. These interventions lasted for about four months with the exception of the intervention of Donelly et al. (2009), which lasted for three years. Davis et al. (2011) speculated that a longer intervention may result in more benefit. Short-term interventions may partly explain the somewhat diverging results. 21 The associations of physical activity and academic achievement have also been measured by longitudinal studies. In the longitudinal study by Stevens et al. (2008), physical activity, but not physical education, was positively associated with mathematics and reading achievement in both boys and girls. Whereas Carlson et al. (2008) reported that girls who had the highest amount of physical education had better math and reading achievement compared to girls with the lowest physical education exposure, while no effect was realized for boys. Haapala et al. (2014a) examined the association of different types of physical activities in first grade with reading and arithmetic skills in grades 1–3. According to the results, children who had more physical activity during recess and children who most often commuted actively to school had better reading fluency across grades 1–3. In addition, children who engaged in organized sports had better arithmetic skills. However, the results were slightly different when data was analysed separately for boys and girls. Boys who had more total physical activity and most often commuted actively to school had better reading fluency and reading comprehension. However, among girls, total physical activity was positively associated with reading fluency and arithmetic skills only for girls whose parents had university level education, while the association was inverse in girls whose parents were less educated (Haapala et al. 2014a). Booth et al. (2014), in turn, examined the associations of objectively measured MVPA at the age of 11 with academic achievement at the age of 11, 13 and 16. MVPA at the age of 11 predicted higher English scores in both sexes at all ages, and higher science scores in females at age of 11 and 16. (Booth et al. 2014). Booth et al. (2014) concluded that MVPA may have a long-term positive influence on academic performance. These results of the longitudinal studies are slightly inconsistent. The authors highlighted factors like intensity and type of physical activity or other factors independent of physical activity such as social development, which may explain the differences (Carlson et al. 2008, Stevens et al. 2008, Booth et al. 2014, Haapala et al. 2014a). Cross-sectional studies determining the relationship of self-reported physical activity and academic achievement (mostly examined by GPA) in children and adolescents have reported a positive association between physical activity and overall academic achievement (Kristjansson et al. 2009, Vindfeld, Schnohr & Niclasen 2009, Fox et al. 2010, Kantomaa et al. 2010, Kristjansson, Sigfusdottir & Allegrante 2010, Edwards, Mauch & Winkelman 2011, So 2012, Kantomaa et al. 2013, Shi et al. 2013). In addition, reported physical activity was especially associated with math performance in girls (Martínez-Gómez et al. 2012), and in boys (O'Dea & Mugridge 2012). However, studies using objectively measured physical activity to determine association with academic achievement have not been as unanimous. According to Kwak et al. (2009), objectively measured vigorous physical activity was positively associated with overall academic achievement in girls, but not in boys. In addition, Telford et al. (2012) reported that objectively measured physical activity was associated with better writing scores, but not with reading or math scores. However, LeBlanc et al. (2012) reported that objectively measured MVPA was not associated with performance gains in academic achievement tests in children. The effects of acute exercise on academic achievement has also been slightly inconsistent. Acute aerobic exercise has been shown to improve reading comprehension (Hillman, Pontifex & Themanson 2009, Duncan & Johnson 2014) and 22 spelling (Duncan & Johnson 2014), have no effect on sentence comprehension (Duncan & Johnson 2014) or spelling and arithmetic (Hillman, Pontifex & Themanson 2009), and attenuate arithmetic (Duncan & Johnson 2014). The timing of the academic testing after an acute bout of exercise may be one factor explaining these conflicting findings. According to Hillman et al. (2009), acute exercise only enhanced reading comprehension, which was measured first, and had no effects on spelling or arithmetic, which were assessed following reading comprehension. This suggests that the benefits of acute exercise may subside over time. Physical fitness has often been used as a proxy indicator of regular physical activity, when the association of physical activity and academic achievement has been examined. In the longitudinal study by London and Castrechini (2011), persistently fit children had higher English and math test scores compared to persistently unfit children. Similarly, Wittberg, Northrup and Cottrell (2010) reported that students who stayed in the “healthy” fitness zone from fifth to seventh grade had significantly higher academic scores than students who stayed in the “needs improvement” zone. According to previous cross-sectional studies, good physical fitness has been associated with better performance in literature and math test scores (Chomitz et al. 2009, Blom et al. 2011, Van Dusen et al. 2011). In addition, aerobic fitness in particular has had a positive association with literature and math test scores (Roberts, Freed & McCarthy 2010, Welk et al. 2010, Wittberg et al. 2010, Davis & Cooper 2011, Scudder et al. 2014). In the study by Edwards, Mauch and Winkelman (2011), higher aerobic fitness was related to math scores, but not reading scores. Moreover, Padilla-Moledo et al. (2012) reported a positive association between muscular fitness and academic achievement. However, there are also studies reporting no associations between physical fitness and academic achievement (Chic & Chen 2011, Wingfield et al. 2011), aerobic fitness and academic skills (Haapala et al. 2014b, Moore et al. 2014), and muscular fitness and math and reading scores (Edwards, Mauch & Winkelman 2011). In addition, in some studies the association between fitness and academic achievement has been different for girls and boys. Eveland-Sayers et al. (2009) reported that aerobic fitness was positively associated with math and reading scores only in girls, while Kwak et al. (2009) reported that cardiovascular fitness was positively associated with GPA only in boys. To sum up the association between physical activity and academic achievement, it seems that children who are more physically active have better academic achievement. However, the reports to date have generally been weak and inconsistent. Especially, there is no intervention, longitudinal, or cross-sectional studies that have replicated the study design and confirmed the results. In addition, previous studies were conducted in various countries with varying educational systems, which makes it difficult to compare the results. Moreover, most of the previous studies have used self-reported measurements of physical activity and academic performance; only a few have objectively measured physical activity to determine the association with educational outcomes. In conclusion, existing studies only show that certain physical activities are associated with certain academic achievements in certain child populations. 23 1.3.2 Physical activity in association with cognitive functions In recent intervention studies, physical activity has been reported to benefit children’s cognitive functions. Ardoy et al. (2014) reported that participating in four high-intensity PE lessons per week improved children’s overall cognitive performance (including non-verbal and verbal abilities, abstract reasoning, spatial ability, numerical ability and verbal reasoning), compared to two or four hours of normal PE lessons per week. In addition, Reed et al. (2010) reported that intervention integrating physical activity into core curricula at school enhanced performance in fluid intelligence, which is considered to illustrate general intelligence and cognitive ability. Likewise, intervention providing 45 minutes of daily physical education led to improvements in fluid intelligence and perceptual speed (Reed et al. 2013). However, the improvements were only observed in certain sections of the fluid intelligence test and were dependent on age and gender (Reed et al. 2013). In particular, executive functions have been reported to benefit from physical activity interventions. Different types of interventions like physical education interventions (Fisher et al. 2011, Spitzer & Hollmann 2013, Crova et al. 2014) and schoolbased physical activity programmes (Davis et al. 2011, Chaddock-Heyman et al. 2013) led to improvements in selective attention and inhibition performance. In addition, aerobically intense physical education intervention and afterschool physical activity programme enhancing aerobic fitness improved children’s spatial working memory and temporal working memory, respectively (Fisher et al. 2011, Kamijo et al. 2011). However, in Puder et al.’s (2011) study, a multidimensional physical activity programme enhancing fitness did not affect spatial working memory or selective attention. In addition, physical education programme including cognitively challenging activities had no effect on verbal working memory (Crova et al. 2014). Puder et al. (2011) speculated that the fairly weak reproducibility of the measures in 5-year-old children and the lack of power may explain contradictory results. Davis et al. (2011) reported that overweight children participating in aerobic exercise programme had higher scores in planning scale assessing strategy generation and application, self-regulation, intentionality and utilization of knowledge, but not in attention, simultaneous or successive sub-scales of the Cognitive Assessment System compared to control children at post-test. They stated that only planning scale measures executive functions, and especially, executive functions are cognitive functions that seems to benefit from physical activity (Davis et al. 2011). However, Fisher et al. (2011), reported that aerobically intense physical education intervention had no effect on any scale of Cognitive Assessment System in healthy children. According to Fisher et al. (2011), intervention was not able to increase time spent in MVPA during physical education lessons enough, which may explain the lack of effect. In addition, the respond to physical activity may differ between lean and overweight children (Davis et al. 2011). In the study Monti, Hillman and Cohen (2012), there were no differences in relational memory performance between children participating in an aerobic exercise programme and control children after a nine-month after-school intervention promoting fitness. However, compared to the control children, children who participated in the aerobic exercise intervention displayed eye-movement patterns indicative of superior relational memory accuracy (Monti, Hillman & Cohen 2012). This 24 indicates that behavioural test may not been sensitive enough to detect small changes in memory performance. To sum up the results of these intervention studies, it seems that physical activity enhances children’s cognitive functions, and especially executive functions. However, there are only a few intervention studies and hardly any randomized controlled trials. In addition, there have been differences in how physical activity is implemented, what cognitive domains have been assessed and how they have been measured, and the ages of the children have also been varied across studies, making it hard to draw any further conclusions. The association of physical activity and cognitive functions has also been measured with cross-sectional studies, which supports the results of the intervention studies. In the study by Ruiz et al. (2010), leisure-time physical activity was associated with better cognitive performance, including verbal, numeric and reasoning abilities in adolescents. In addition, Castelli et al. (2011) reported that engagement in vigorous physical activities had a positive association with performance in inhibitory control task. Besides chronic effects of physical activity, recent studies have also shown that acute physical exercise induces positive changes in free-recall memory performance (Pesce et al. 2009), working memory performance (Hill et al. 2010), and inhibitory control performance (Budde et al. 2008, Hillman, Pontifex & Themanson 2009, Hill et al. 2010, Best 2012, Drollette et al. 2012, Gallotta et al. 2012, Hogan et al. 2013). In addition, in the study by Drollette et al. (2014), the effects of acute aerobic exercise on inhibitory control performance were examined in two groups of children categorized as higher- and lower-performers. Children were divided into two groups according to their inhibitory control performance following the resting session. According to the results, higher-performers maintained their performance, while lower-performers improved their performance in inhibitory control task following exercise. However, acute physical activities have not improved spatial working memory performance (Drollette et al. 2012) or inhibitory control (Stroth et al. 2009) in all studies. According to Stroth et al. (2009), the selected task assessing inhibition was too easy for children, and due to a ceiling effect the effects of acute exercise on inhibition was not revealed. Drollette et al. (2012), in turn, suggested that acute exercise may affect selectively to executive functions enhancing inhibitory control, but not working memory. As stated before, physical fitness has been used as a proxy measure of regular physical activity. Associations of physical fitness and cognition have been measured with cross-sectional and longitudinal studies. According to the longitudinal study by Chaddock et al. (2012b), children with higher aerobic fitness outperformed less fit children in an inhibitory control task at the initial time of fitness testing, as well as one year later. In addition, children with high aerobic fitness have demonstrated better inhibitory control performance in cross-sectional studies (Buck, Hillman & Castelli 2008, Hillman et al. 2009, Chaddock et al. 2010b, Pontifex et al. 2011, Voss et al. 2011, Wu et al. 2011, Chaddock et al. 2012a, Pontifex et al. 2012, Crova et al. 2014). In addition, Hogan et al. (2013) reported a positive interactive effect of physical fitness level and acute aerobic exercise on inhibitory control: higher fit children had shorter reaction times after exercise compared to a resting condition. However, according to Stroth et al. (2009) and Castelli et al. (2011), aerobic fitness were not associated with inhibitory control. 25 High aerobic fitness have also been connected to better memory performance (Chaddock et al. 2010a, Chaddock et al. 2011, Raine et al. 2013), planning ability (Davis & Cooper 2011), attentional performance (Davis & Cooper 2011, Wu & Hillman 2013) and arithmetic cognition (Moore et al. 2014). Furthermore, Chaddock et al. (2012c) reported that aerobically more fit children outperformed less fit children in a virtual street-crossing task. More fit children maintained street-crossing performance when distracted, whereas the performance of less fit children attenuated when conversing on a phone. Finally, Åberg et al. (2009) reported that cardiovascular fitness was associated with intelligence in young adulthood. Still, fitness has not always been reported to have an association with cognition. Ruiz et al. (2010) reported that neither aerobic nor muscular fitness was associated with overall cognitive performance. Whereas Davis et al. (2011) reported that fitness was not associated with the Simultaneous (processing with spatial and logical questions) or Successive (analysis/recall of stimuli arranged in sequence) sub-categories of the Cognitive Assessment System. These apparent inconsistencies in associations between physical activity and cognition suggest that physical activity may selectively affect certain cognitive functions such as executive functions. However, it seems that studies have most often measured the association between physical activity and executive functions. In addition, there are no studies with exactly the same design, which attenuates the interpretation of the results. Furthermore, physical fitness has often been used as a proxy indicator of regular physical activity, without direct measurement of actual physical activity levels. This is particularly important because in childhood habitual physical activity is rarely intensive and lengthy enough to enhance aerobic fitness, and therefore, the relationship between physical activity and fitness may not be meaningful (Armstrong, Tomkinson & Ekelund 2011). This may cause discrepancies in results. For example, Ruiz et al. (2010) and Castelli et al. (2011) reported that physical activity was associated with cognition, while fitness was not. In summary, it seems that physical activity benefits children’s cognitive functions. However, evidence of the favourable effects of physical activity on cognitive functions in healthy children and adolescents is still somewhat inconsistent and based on scarce research data. Especially, the type of physical activity assessed and measurements of physical activity used have varied across studies. Similarly, the assessments of certain cognitive domains have been divergent. Moreover, only a few studies have measured a broad range of cognitive functions. This highlights the need for new studies to clarify the benefits of physical activity on different dimensions of cognitive functions. 26 1.4 Associations of sedentary behaviour, academic performance and cognitive functions Since the invention of television, the effects of children’s exposure to TV has been discussed in terms of academic achievement and cognitive development (Maccoby 1951). According to these early studies, excessive television viewing may have an unfavourable impact on children’s cognitive functions and academic achievement, but the evidence is ambiguous (Gaddy 1986, Gortmaker et al. 1990, Levine & Waite 2000), suggesting that the effects of TV consumption may vary as a function of socioeconomic background, intelligence and other sub-groups of children (Anderson & Collins 1988, Beentjes & Van der Voort, Tom HA 1988, Smith 1992, Cooper et al. 1999). In addition, the association may not be linear, implying that the association turns negative with high amounts of TV viewing (Williams et al. 1982). Newer studies seem to support this observation by reporting an inverse association between high amounts of TV viewing and lower academic achievement (Chernin & Linebarger 2005, Schmidt & Vandewater 2008, Tremblay et al. 2011b), curvelinearity of the association (Razel 2001) and an association between high amounts of TV viewing and attention problems in school-aged children (Schmidt & Vandewater 2008). Especially early-childhood TV exposure may have a negative impact on children’s development, including language, cognition and attention capacity, and readiness to attend school (Clarke & Kurtz-Costes 1997, Wright et al. 2001, Christakis 2009). In addition, early-childhood TV viewing has been associated with attention problems (Christakis et al. 2004, Zimmerman & Christakis 2007), attenuated cognitive outcomes (Zimmerman & Christakis 2005) and decreased academic achievements (Pagani et al. 2010) at school-age. However, inconsistent results have also been reported in early childhood, showing no association between TV exposure and language or visual motor skills (Schmidt et al. 2009). On the other hand, educational TV viewing has been linked to enhanced academic and cognitive outcomes (Linebarger et al. 2004, Chernin & Linebarger 2005, Kirkorian, Wartella & Anderson 2008, Schmidt & Vandewater 2008). Besides TV viewing, computer use, video game playing and other screenbased sedentary behaviour have also been linked to cognitive skills and academic achievement. According to recent studies, playing computer or video games may even benefit cognitive functions, especially attentional, visuospatial and problemsolving skills in young adults (Spence & Feng 2010, Granic, Lobel & Engels 2013), as well as in children and adolescents (Subrahmanyam et al. 2001, Schmidt & Vandewater 2008). In addition, computer use has been associated with slightly better academic performance (Subrahmanyam et al. 2001). However, previous results have not been consistent. The associations of screen-based sedentary behaviour, academic achievement and cognitive functions in school-aged children are described more in detail below. 27 1.4.1 Sedentary behaviour in association with academic achievement According to recent studies, media use in childhood – especially time spent viewing TV, playing video games, and using the Internet – has a negative association with academic achievement. In a longitudinal study by Sharif, Wills and Sargent (2010), screen-time exposure – including TV viewing, video game playing and the presence of a television in the bedroom – had adverse effects on improvements in overall school performance in children aged 10–14 years. Similarly, Mößle et al. (2010) reported the results of two studies among primary school students. According to the cross-sectional results among fourth-grade students, the time used for playing computer or video games and watching TV, DVDs and videos was negatively associated with school achievement, including grades in mother tongue, science and math (Mößle et al. 2010). Similarly, in another study that evaluated cross-sectional associations, media use – especially playing computer games – was negatively associated with marks in mother tongue, foreign language and science, but to a lesser extent with marks in math at all measurement occasions in 3rd, 4th and 5th grades. In addition, in longitudinal analysis, negative correlations were observed between the duration of daily computer game playing and academic achievement, while the duration of TV usage had only few negative associations with academic achievements (Mößle et al. 2010). Especially, frequent TV viewing in childhood and adolescence has been connected to poor reading achievement in childhood (Ennemoser & Schneider 2007), poor academic grades, failure to complete high school, negative attitudes towards school and poor homework completion in adolescence, as well as long-term academic failure (Johnson et al. 2007) and poor educational achievement at the age of 26 (Hancox, Milne & Poulton 2005). In cross-sectional studies, high amounts of television viewing have been shown to have negative associations with math and reading achievement in 6- to 13-year-old children (Shin 2004) and school performance in 5th–8th graders (Sharif & Sargent 2006). Furthermore, children having a TV in their bedroom, the number of television sets at home and the number of hours that televisions are on have been negatively associated with academic achievement (Gentile & Walsh 2002, Borzekowski & Robinson 2005, Espinoza 2009). Munasib and Bhattacharya (2010), however, reported no association between television viewing and academic achievement in 5–10 years old children after adjusting for socioeconomic determinants, parents’ TV-viewing behaviour and parents’ role in monitoring children’s viewing. Similarly, in a longitudinal study by Bittman et al. (2011), in which two age cohorts (younger cohort aged 0–1 years and older cohort aged 4–5 years at the beginning of the study) were followed for four years, television viewing was not associated with language skills after the parents’ involvement in the child’s media use was taken into account. In addition, Haapala et al. (2014a) reported that television viewing in first grade was not associated with reading and arithmetic skills in grades 1–3. Besides TV viewing, computer use and video game playing have been reported to have negative associations with academic achievement. However, these results are more inconsistent: in the intervention study by Weis and Cerankosky (2010), the effects of video game ownership on academic achievement were evaluated in boys aged six to nine. Boys were randomly assigned into two groups: the experi28 mental group received a video game system at the beginning of the four month intervention, whereas the control group received a video game system after follow-up assessments. According to the results at the post-test, children with video games spent more time playing them and received lower reading and writing scores, but not math scores, compared to control children (Weis & Cerankosky 2010). In a longitudinal study by Jackson et al. (2011), video game playing was associated with lower academic achievement, especially lower GPAs in 12-year-old children. Similarly, according to the study by Bittman et al. (2011) presented earlier, the ownership of game consoles was negatively associated with literature achievements in the older cohort (children aged 8 years). However, in the study by Sharif and Sargent (2006), video game use was not associated with school performance in 5–8 graders. According to Ferguson et al. (2013), violent video game exposure had neither positive nor negative predictive short-term or long-term associations with math achievement in children and youths aged 10–17 years. Similarly, in the longitudinal study by Willoughby (2008), the association between computer game play and academic performance was not significant at the age of 14–16. Haapala et al. (2014a), in turn, reported that high levels of computer use and video game playing in first grade predicted better arithmetic skills among boys in grades 1–3, while no effect was realized for girls. According to Jackson et al. (2011), Internet use was associated with better reading skills in 12-year-olds, but only in children with initially low reading skills. The same kind of effect was not realized for math skills (Jackson et al. 2011). Moreover, moderate use of the Internet has been connected to more positive academic performance than non-use or high use in ninth to 12th graders (Willoughby 2008, Kim & So 2012). In the other study by Jackson et al. (2006), those children and youths from low-income families who used the Internet more frequently achieved higher reading (but not math) test scores and higher grade point averages 6, 12 and 16 months later, compared to children who used the Internet less frequently. In addition, Bittman et al. (2011) reported that computer use was associated with higher developed language skills. Likewise, Borzekowski and Robinson (2005) reported that computer access and use were positively associated with academic achievement. In conclusion, the association between screen-based sedentary behaviour and academic achievement is still somewhat inconsistent. The association seems to depend on the type of screen-based behaviour assessed, the age of the children, and other factors like socioeconomic status. For instance, media use may be more controlled in small children, and the effects of screen time on academic achievement may become emphasized in older children (Gortmaker et al. 1990). In addition, the amount of time spent in front of the screen (non-use vs. moderate use vs. high use) and content of screen time may be critical factors affecting the association between screen time and academic achievement. Screen time – such as playing video games or using the Internet – may be beneficial when the use is light or moderate, but may be harmful when the use is excessive. However, the literature concerning screenbased sedentary time in association with academic achievement is divergent and there are no replicated study designs, so more research is needed to clarify the association. Besides, to our knowledge, no one has studied the association between objectively measured sedentary time and academic performance. 29 1.4.2 Sedentary behaviour in association with cognitive functions Extensive screen time has been linked to an elevated risk of attention and learning difficulties. In the study by Swing et al. (2010), high amounts of screen time were associated with attention problems. The association of screen time and attention problems was similar for both TV viewing and video game playing, as well as for both age groups (6–12 years and 18–32 years) (Swing et al. 2010). In addition, a high amount of TV viewing in both childhood and adolescence has been associated with frequent attention difficulties in adolescence. (Johnson et al. 2007, Landhuis et al. 2007). However, in the intervention study by Weis and Cerankosky (2010), where 6–9-year-old boys were randomly assigned to an experimental group receiving a video game system at the beginning of the four-month intervention or a control group receiving a video game system after the follow-up, the ownership and playing of video games did not affect attention problems. On the other hand, Weis and Cerankosky (2010) observed higher learning problems in boys with the video game system compared to control childen. Excessive screen time has also been connected to weaker executive functions in some studies, but not in the others. Mizuno et al. (2013) reported that high amounts of television viewing were associated with a weaker ability for adolescents to divide attention (mean age 13 years). In the study by Dye, Green and Bavelier (2009), the inhibitory control skills of action-game players and a non-playing control group aged 7 to 22 years were compared. The results, however, showed that action-game playing was associated with enhanced attention skills (Dye et al. 2009). Video game playing has been linked to decreased verbal memory performance (Dworak et al. 2007). Drowak et al. (2007) studied the effects of excessive television and video game exposure on the visuospatial and verbal memory performance of 13-year-old children. According to the results, excessive video game playing, but not television viewing, decreased verbal memory performance compared to the basal condition. Visuospatial memory performance was not affected by either television or video game exposure (Dworak et al. 2007). In turn, Ferguson et al. (2013) reported that violent video game exposure had neither positive nor negative predictive short-term or long-term association with visuospatial cognition in children and youths aged 10–17 years. In addition, in the study of Ruiz et al. (2010), television viewing and video game playing were not associated with overall cognitive performance, including verbal, numeric and reasoning abilities in adolescents. Differing research results based on scarce and heterogeneous research data, indicates that the association between sedentary behaviour and cognition is more complicated than previously believed and needs clarification. In addition, to our knowledge, no previous studies have examined the associations of objectively measured overall sedentary time on cognitive functions in children. 30 1.5 Summary of literature In sum, children today spend excessive amounts of time engaged in sedentary activities and not enough time in physical activities. Low levels of physical activity have raised concerns over the effects of a physically inactive lifestyle on children’s physical health and, recently, also on children’s learning. Learning is an active and interactive process, which results in changes in skills, knowledge and behaviour. At school, academic achievements have been used to monitor and evaluate children’s learning. Cognitive functions, especially executive functions, can be seen as prerequisites of learning. According to recent studies, physical activity may benefit children’s cognitive functions and academic performance, while sedentary behaviour, especially screen-based sedentary behaviour, may attenuate them. Although the number of studies examining the association of physical activity and sedentary behaviour with academic and cognitive performance has almost doubled during recent years, research in this area is still in its infancy, and the evidence is somewhat inconsistent. In particular, the definitions, patterns and measurements of cognitive functions, academic performance and physical activity have varied across different studies. In previous studies, physical activity has often been measured with self-reports or physical fitness has been used as a proxy measure of regular physical activity instead of measuring physical activity levels directly. Likewise, sedentary behaviour has usually been assessed with self-reports of screen time and, especially in earlier studies, with self-reports of the time spent viewing TV. Furthermore, the use of computerized test batteries to measure cognitive functions has increased worldwide, but the psychometric properties of such test batteries have not been adequately measured. Future studies are needed to clarify the associations between physical activity, sedentary behaviour, academic performance and cognitive functions in school-aged children. 31 32 2 AIMS OF THE STUDY The purpose of the present thesis was to determine how physical activity and sedentary behaviour are associated with academic performance and cognitive functions in elementary school-aged children. The specific aims were: Aim 1. To examine the associations of self-reported and objectively measured physical activity and sedentary behaviour with teacher-rated academic performance in children. Aim 2. To evaluate the internal consistency and the one-year stability of seven tests of the Cambridge Neuropsychological Test Automated Battery (CANTAB) used to measure visual memory, executive function, and attention in children. Aim 3. To examine how objectively measured and self-reported physical activity and sedentary behaviour are associated with cognitive functions in children. 33 34 3 MATERIAL AND METHODS 3.1 Study population and data collection During spring 2011, 475 5th and 6th graders from five schools in the Jyväskylä school district in Finland were invited to participate in the study, which included a self-reported questionnaire filled out in the classroom, an objective measurement of physical activity for seven days, and cognitive tests. Fifty-eight percent (N=277) of 475 eligible children participated in the study. Children engaged in normal curriculum-based instruction, and the language of instruction was Finnish. They had normal or corrected-to-normal vision. Of the 277 children, 230 were selected to participate in cognitive tests according to successful objective measurement of their physical activity. If a child’s physical activity measurement did not succeed, because of technical problems or the child did not remember to wear the accelerometer, they were not invited to the cognitive tests. Seven children (three boys and four girls) were excluded from the analysis of the association of physical activity and cognitive functions because, according to their parents’ survey, they had physical disabilities, chronic diseases or severe learning disabilities. During spring 2012, students who had been fifth graders in spring 2011 were invited to participate in follow-up measurements. Seventy-four children (49% of 151 eligible) participated in these followup measurements. The sample characteristics of the study are presented in Tables 1–3. 35 TABLE 1. Sample characteristics concerning explanatory variables. Boys Explanatory variables Physical activity Self-reported MVPA (d/ week with ≥60 min MVPA) Objectively measured MVPA (min/day) Sedentary behaviour Self-reported screen time (h/day) TV Computer/video games Computer use (other than playing) Objectively measured sedentary time (%/day) Girls pa All Mean±SD N Mean±SD N Mean±SD N 5.2±1.8 121 4.9±1.6 153 5.0±1.7 274 0.049 59.9±22.3 95 56.3±17.1 125 57.9±19.5 220 0.623 3.8±2.0 121 3.5±1.9 154 3.6±1.9 275 0.095 1.6±1.0 122 1.6±1.0 154 1.6±1.0 276 0.663 1.3±0.9 121 0.7±0.8 154 1.0±0.9 275 <0.001 0.9±0.7 122 1.2±0.9 154 1.1±0.8 276 0.011 39.6±3.5 95 40.8±3.1 125 40.3±3.3 220 0.006 Abbreviations: SD, standard deviation; MVPA, moderate to vigorous physical activity. a p-values for the gender differences. TABLE 2. Sample characteristics concerning outcome variables. Boys Outcome variables Academic performance Grade point average (range 4– 10) Cognitive function Visual memory PRM no. of correct responses (max 24) SRM no. of correct responses (max 20) Executive functions SSP span length (max 9) SOC no. of problems solved in minimum moves (max 12) IED no. of children who completed the test (%) Attention RTI five-choice movement time (ms) RTI five-choice reaction time (ms) RVP A’ (range 0–1) Girls pa All Mean±SD N Mean±SD N Mean±SD N 8.1±0.7 122 8.4±0.6 153 8.2±0.7 275 <0.001 20.9±2.80 99 20.8±2.3 131 20.9±2.5 230 0.478 16.5±1.6 99 16.8±1.7 131 16.7±1.7 230 0.152 6.5±1.3 99 6.7±1.3 131 6.6±1.3 230 0.385 7.7±1.8 99 7.6±1.7 131 7.6±1.8 230 0.746 67 99 67 131 67 230 0.935 329±74 99 364±87 131 349±83 230 0.001 300±36 99 318±32 131 310±35 230 <0.001 0.97±0.02 99 0.97±0.03 131 0.97±0.02 230 0.973 Abbreviations: SD, standard deviation; PRM, Pattern Recognition Memory; SRM, Spatial Recognition Memory; SSP, Spatial Span; SOC, Stockings of Cambridge; RTI, Reaction Time; RVP, Rapid Visual Information Processing; IED, Intra-Extra Dimensional Set Shift. a p-values for the gender differences. TABLE 3. Sample characteristics concerning other variables. Boys Girls pa All Other variables Mean±SD N Mean±SD N Mean±SD N Age (years) Families in which the highest level of parental education was tertiary-level education (%) Family income (€) Parents, who are married or cohabiting (%) Children with learning difficulties (%) Children with need for remedial education Amount of sleep (h) 12.2±0.7 123 12.2±0.6 154 12.2±0.6 277 0.765 80 94 79 126 79 220 0.826 65319±29443 69 63755±27346 103 64383±28132 172 0.722 77 94 75 126 76 220 0.734 9 91 6 124 7 215 0.371 19 92 14 126 16 218 0.405 9.1±0.8 122 9.0±0.7 154 9.1±0.7 276 0.414 18.9±3.4 121 18.9±3.2 149 18.9±3.3 270 0.997 Body mass index Abbreviations: SD, standard deviation. a p-values for gender differences. 3.2 Measurements 3.2.1 Self-reported physical activity and screen time Physical activity and screen time were assessed with a self-reported questionnaire used earlier in the World Health Organization (WHO) Health Behaviour in Schoolaged Children (HBSC) study (Currie et al. 2012). Self-reported MVPA was measured with the following question: “Over the past 7 days, on how many days were you physically active for a total of at least 60 minutes per day?” The response categories were: 0 days, 1 day, 2 days, … 7 days. There was a short description about what kind of physical activity should be taken into account when answering the question: “In the next question, physical activity is defined as any activity that increases your heart rate and makes you get out of breath some of the time when you are for example exercising, playing with friends, commuting actively to school or in physical education lessons.” Examples included running, walking quickly, rollerblading, biking, dancing, skateboarding, swimming, snowboarding, cross-country skiing, soccer, basketball, and Finnish baseball. Test-retest agreement for self-reported MVPA has been very good (ICC=0.82) (Booth et al. 2001, Liu et al. 2010). Self-reported screen time was evaluated with the question: “About how many hours a day do you usually a) watch television (including videos), b) play computer or video games, or c) use a computer (for purposes other than playing games, for example, emailing, chatting, or surfing the Internet or doing homework) in your free time?” The response options were: not at all, about half an hour per day, about an hour a day, about two hours per day, … about five hours per day or more. Children responded separately for both weekdays and weekends. Test-retest agreement for watching television (ICC=0.72–0.74) and for playing computer or video games (ICC=0.54–0.69) has been substantial, and fair to moderate (ICC=0.33–0.50) for using the computer (Liu et al. 2010). Daily screen-time averages were calculated by adding these three questions together, including weekdays and weekends. 3.2.2 Objective measures of physical activity and sedentary time Children’s physical activity was measured objectively by using the ActiGraph GT1M/GT3X accelerometer with one vertical axle. Children wore the accelerometer on their right hip with an elastic waistband during waking hours for seven consecutive days. During bathing, swimming, and other water activities, it was requested that the monitor be removed, because it was not water-resistant. The ActiLife accelerometer software (ActiLife version 5; http://support.theactigraph.com/dl/ActiLife-software) was used to initialize the monitors and download the data. Epoch length was 10 seconds and non-wearing time 30 minutes. Customized software was used for data reduction and analysis. A cut-off value of 2,296 counts per minute was used for MVPA (Evenson et al. 2006) and 100 counts per minute for sedentary time. Children were included in the analysis if they had valid data for at least 500 minutes per day on two weekdays and on one weekend day. In order to compare children who had worn the accelerometers for different amounts of time per day, objectively measured sedentary time was expressed as the percentage of daily amount of time in which data was being registered. 39 3.2.3 Academic performance Academic achievement scores (grades in individual school subjects and GPAs) were provided by the education services of the city of Jyväskylä. Individual grades were assessed in the following school subjects: mother tongue (in most cases Finnish or Swedish), first foreign language (started in 3rd grade), mathematics, physics/chemistry, biology, history, geography, religion or ethics, visual arts, music and physical education. The grades refer to numerical assessment on a scale of 4–10, where 4 denotes a failure (US grade: F) and 10 denotes excellent knowledge and skills (US grade: A). The GPAs were calculated on the basis of the individual grades and were used as a measure of academic achievement in the analysis. A Finnish GPA 5.0–5.9 equals 1.0 in US GPA, 6.0–6.9 equals 2.0, 7.0–8.9 equals 3.0, and 9.0–10.0 equals 4.0, respectively. 3.2.4 Cognitive functions CANTAB (CANTABeclipse version 3 on a PaceBlade Slimbook P110 tablet PC with a 12-inch touch-screen monitor and Windows XP Professional operating system) was used to assess a broad range of cognitive functions: a) visual memory (Pattern Recognition Memory [PRM] and Spatial Recognition Memory [SRM]), b) executive function (Spatial Span [SSP], Stockings of Cambridge [SOC], Intra-Extra Dimensional Set Shift [IED]), and c) attention (Reaction Time [RTI] and Rapid Visual Information Processing [RVP]) (Table 4). The tests were run individually with the help of a trained research assistant and according to the standard protocol. Standard instructions for the tests were provided in the CANTAB manual and were translated into Finnish. The execution required about 45 minutes. The test battery was administered in a silent room without distractions. A Motor Screening Task measuring simple psychomotor speed and accuracy was used as a training procedure at the beginning of a test session. A more detailed description of the tests can be found in the original study II. TABLE 4. Summary of the CANTAB tests used to measure different dimensions of cognitive function. Dimension of cognitive function Test Visual memory Pattern Recognition Memory Spatial Recognition Memory Spatial Span Stockings of Cambridge Intra-Extra Dimensional Set Shift Reaction Time Rapid Visual Information Processing Executive function Attention 40 Abbreviation PRM SRM SSP SOC IED RTI RVP 3.2.4.1 Visual memory Visual memory performance was assessed with PRM and SRM. PRM measures recognition memory for visual patterns and SRM for spatial locations in a two-alternative forced-choice paradigm. In these tests, children had to remember presented geometric patterns (PRM) or the locations of white squares (SRM) and discriminate them from novel patterns and locations. The scores in these tasks are based on the number of correct responses (PRM maximum 24, SRM maximum 20). 3.2.4.2 Executive function Children’s executive functions were assessed with SSP, SOC and IED tests. The SSP is based on the Corsi Blocks task (Milner 1971), which measures the length of visuospatial memory span. In this test, a specified number of white boxes changed their colour one by one and children had to reproduce the same sequence by touching the boxes in the same order that the boxes had changed colour. The score in the task is based on the length of the maximum sequence that the child can reproduce. The SOC is a computerized version of the Tower of London task (Shallice & Shallice 1982, Owen et al. 1990), measuring spatial planning and spatial working memory. In this test, children had to move coloured balls in the lower part of the screen to the same position in the stockings as they were in the upper part of the screen. Children had a specified number of moves to use. The score in this task is based on the number of problems that the child solves with a minimum number of moves. IED is a computerized analogue of the Wisconsin Card Sorting test and measures rule acquisition and reversal in a set-shifting condition. Specifically, it measures the ability to focus attention on different stimuli within a relevant dimension and shift attention to a previously irrelevant dimension. In this task, there are nine stages with increasing levels of difficulty. The children were instructed to choose one of two different dimensions: one was correct and the other was incorrect. According to immediate feedback, they were expected to choose the correct pattern and learn the rule. The score in this task is based on the number of stages completed. 3.2.4.3 Attention The children’s attention abilities were assessed with RTI and RVP. RTI measures the children’s speed of response to an unpredictable visual target. In the unpredictable five-choice scenario, a yellow spot appeared randomly in one of five circles on the screen, and children were asked to respond as soon as possible by touching the correct circle. The scores in this task are based on reaction time (ms) and movement time (ms). Measuring sustained attention, the RVP is similar to the Continuous Performance Task. In this test, children had to press a button every time they discriminated the target sequence (digits 3, 5, and 7) from the digits appearing in a pseudorandom order at the rate of 100 digits per minute. The score in this task is based on RVP A’, which measures how successful the child is at detecting target sequences (range: .00 to 1.00; bad to good). 41 3.2.5 Potential confounders The parent or the child’s main caregiver filled in a questionnaire concerning family background. The mother’s and father’s education were measured with the following questions: “What is the highest level of mother’s/father’s education?” The highest level of parental education (calculated on the basis of the mother’s and father’s education) were categorized as 1) tertiary-level education and 0) basic or upper secondary education. Family income was measured with the question: “What was your household's total income last year (with taxes)?” Marital status was assessed with the question: “Which of the following best describes the marital status of the child's main caregiver? A) Married or cohabiting with child’s mother/father, B) Divorced, single parent, C) Divorced, joint custody, D) Divorced, a new marriage, E) Single, D) Widow.” The marital status of the main caregiver was categorized as 1) married or cohabiting with child’s mother/father and 0) divorced or single/widow. Children’s learning difficulties and need for remedial education were measured with the questions: “Has your child been diagnosed with a learning difficulty?” and “Has your child received remedial education?” Children’s learning difficulties and need for remedial education were categorized as 1) yes or 0) no or don’t know. Children also reported times when they usually went to sleep and woke up on school days. 3.3 Ethics statement The study was approved by the Ethics Committee of the University of Jyväskylä, and it followed the principles of the Declaration of Helsinki and the Finnish legislation. Participation in the study was voluntary, and all participants had the right to drop out of the study at any time without any specific reason. Only children with a fully completed consent form (Certificate of Consent signed by a parent/guardian and the child) on the day of the first measurements were included in the study. 3.4 Analytical strategies 3.4.1 Study I In Study I, the main objective was to examine cross-sectional associations of physical activity and sedentary behaviour with academic achievement. In addition, potential confounding factors were taken into account. These associations were examined with analysis of variance and linear regression analysis using the SPSS 19.0 for Windows statistical package (SPSS (2010) IBM SPSS Statistics 19 Core System User’s Guide (SPSS Inc., Chicago, IL)). For the analysis of variance, children were divided into tertile groups (33% each) according to the amount of objectively measured MVPA (first tertile ≤47.0 min, second tertile 47.1–65.0 min, third tertile ≥65.1 min) and sedentary time (first tertile ≤38.4%, second tertile 38.5–41.4%, third tertile ≥41.5%). In addition, children were classified into groups according to the self-reported MVPA (1=0–2 days/week, 2=3–4 days/week, 3=5–6 days/week, 4=7 days/week) and screen time (1=0.00– 42 1.99 h/day, 2=2.00–2.99 h/day, 3=3.00–3.99 h/day, 4=4.00–4.99 h/day, 5=≥5.00 h/day). Before proceeding with multiple regression, the Pearson’s correlation coefficients for continuous variables were calculated to estimate associations between single variables and GPA. To investigate whether the associations between self-reported or objectively measured MVPA and GPA are quadratic, quadratic terms were calculated using an equation x2=((x-mean(x))*(x-mean(x)), where x was logarithmically transformed, self-reported MVPA or logarithmically transformed, objectively measured MVPA. After that, the enter method for the multiple regression was used. To calculate change in R square, the variables of interest were added to the second block one by one, while all other variables of the model (potential confounders and other variables of interest) were added to the first block. The change in R square for all variables of interest was calculated and tested for significance. In order to study whether the assumptions of the regression analysis were fulfilled, we examined the distribution of model residuals. 3.4.2 Study II In Study II, the main objective was to examine internal consistency and one-year stability of the seven CANTAB tests. The analyses were performed using the SPSS 19.0 for Windows statistical package (SPSS (2010) IBM SPSS Statistics 19 Core System User’s Guide (SPSS Inc., Chicago, IL)) and the Mplus statistical package (Version 7; Muthèn & Muthèn, 1998–2012). The internal consistency of each nonhampering test was estimated with Cronbach’s alpha reliability coefficient (α). The reliability could not be determined for hampering tests. As preliminary analysis of stability, Pearson’s correlation for continuous variables and tetrachoric correlations for dichotomously scored variables were calculated between the assessments. To examine the stability of the CANTAB tests, structural equation models were applied. To combine a very large number of measured variables for each latent factor, item parcels were constructed by summing every third pattern into the same parcel (Little et al. 2002). Item parcels were used in order to achieve continuous indicators and not to end up with too large of a model in terms of the ratio of sample size to number of free parameters (Herzog & Boomsma 2009, Christopher Westland 2010). Underlying latent traits were assumed to be unidimensional. The constructed parcels were then used as indicator variables in the confirmatory factor analyses. The measurement models were first specified at both measurement times to test the association between the observed variables and the underlying factors. After demonstrating the fit of the measurement models, longitudinal confirmatory factor analyses were performed. The baseline stability model was estimated, in which the factor (or factors) in the second assessment (2012) was predicted by the factor (or factors) in the previous measurement point (2011). To detect time invariance in the latent constructs, equality constraints were imposed on the corresponding factor loadings across two time points. Furthermore, if the invariance assumption of a stability model was supported, a more parsimonious model in which all the factor loadings were fixed to be one was estimated. 43 Full information maximum likelihood (FIML) estimation with robust standard errors (MLR) was used under the assumption of data missing at random. Item response theory (IRT) modelling using FIML estimation was applied to the CANTAB tests with hampering nature. The Satorra–Bentler scaled χ2-test, the comparative fit index (CFI), the Tucker–Lewis Index (TLI), the root mean square error of approximation (RMSEA) and the standardized root mean square residual (SRMR) were used to evaluate the goodness of fit of the models. The model fits the data well if the pvalue for the χ2-test is non-significant, CFI and TLI values are close to 0.95, the RMSEA value is below 0.06 and the SRMR value is below 0.08 (Hu & Bentler 1999). A Satorra–Bentler scaled χ2 difference test was conducted for the nested models. If the χ2-test produces a non-significant loss of fit for the constrained model as compared to the baseline stability model, the equality constraints are supported. 3.4.3 Study III In Study III, the main focus was to examine the associations of objectively and subjectively measured physical activity, sedentary behaviour and cognitive tests. On the basis of the reliability and stability results of the CANTAB tests, five of seven cognitive tests were chosen. In addition, structural equation modelling was chosen to examine the associations, because it enables estimation of measurement errors and therefore increases the reliability of cognitive tests. The analyses were performed using the SPSS 19.0 for Windows statistical package (SPSS (2010) IBM SPSS Statistics 19 Core System User’s Guide (SPSS Inc., Chicago, IL)) and the Mplus statistical package (Version 7; Muthèn & Muthèn, 1998–2012). In this study also, item parcels (Little et al. 2002) were constructed and then used as indicators for latent variables. Single scores of every problem or level of the cognitive tests were applied instead of the total score. When the outcome was dichotomous, multiple logistic regression was used to examine physical activity and sedentary time in association with cognitive function. Full information maximum likelihood estimation with robust standard errors was used under the assumption of data missing at random. Gender, the highest level of parental education and the child’s need for remedial education were chosen to represent different aspects of potential confounders and were added to the main analysis. In order to avoid multicollinearity, highly correlated objectively measured MVPA and sedentary time were added to the model using a Cholesky factoring of the predictors (de Jong 1999). The Satorra-Bentler scaled χ2-test, the CFI, the TLI, the RMSEA and the SRMR were used to evaluate the goodness of fit of the models, as described above. 44 4 AN OVERVIEW AND THE MAIN RESULTS OF THE ORIGINAL STUDIES 4.1 Study I: Physical activity, sedentary behaviour, and academic performance in Finnish children Purpose of the study. Study I aimed to examine how both objectively measured and self-reported physical activity and sedentary behaviour are associated with teacherrated academic achievement in children. Based on the previous studies, it was hypothesized that physical activity would have a positive association, while sedentary behaviour would have a negative association with academic performance. Methods. For this purpose, information on children’s academic achievement (GPA) was provided by the education services of the city of Jyväskylä. Children selfreported how many days per week they were physically active (for a total of at least 60 minutes per day) and how many hours per day they spent watching TV, playing video games or computer games, or using a computer for other purposes than play. Children’s physical activity and sedentary time were measured objectively by using an ActiGraph GT1M/GT3X accelerometer. Cross-sectional associations were examined using analysis of variance and linear regression analysis. Results. Our results showed that objectively measured MVPA (p = 0.955) and sedentary time (p = 0.285) were not associated with GPA (Figure 2). However, selfreported MVPA had an inverse U-shaped curvilinear association with GPA (p = 0.001), and screen time had a linear negative association with GPA (p = 0.002), after adjusting for gender, children’s learning difficulties, the highest level of parental education, and amount of sleep (Table 5). Children who were physically active at least 60 minutes per day 5–6 days per week had the highest GPAs, whereas children who were physically active 0–2 days per week had the lowest GPAs (Figure 2). Children who had less than 2 hours per day of screen time had the highest GPAs, whereas children who had more than 5 hours per day of screen time had the lowest GPAs (Figure 2). Conclusions. Our finding that self-reported physical activity was directly, and screen time inversely, associated with academic achievement, is consistent with earlier studies (Fox et al., 2010; Kantomaa et al., 2010; Möble et al., 2010; Sharif et al., 2006). However, objectively measured physical activity and sedentary time were not associated with academic achievement. This inconsistency between the subjective and objective measurements suggests that objective and subjective measurements may reflect different constructs and contexts of physical activity and sedentary behaviour in association with academic outcomes. 45 FIGURE 2. Grade point averages according to a) self-reported MVPA, b) self-reported screen time, c) objectively measured MVPA, and d) objectively measured sedentary time. Results from the regression analysis. 46 TABLE 5. Regression analysis of grade point average (GPA). Academic achievement (GPA) Pearson’s correlations Variables in the model Children’s learning difficultiesa Highest level of parental educationb Amount of sleep (h/day) Gender (female) c Self-reported MVPA Quadratic self-reported MVPAd Self-reported screen time (h/day)e The regression model B (SE) -0.697*** (0.139) Beta (SE) -0.297*** (0.139) 0.247*** 0.307** (0.087) 0.207** (0.087) 0.211*** 0.114* (0.052) 0.129* (0.052) 0.223*** 0.176* (0.072) 0.144* (0.072) 0.003 0.065 (0.061) 0.067 (0.061) -0.368*** ∆R2f 0.004 -0.247*** -0.337** (0.104) -0.199** (0.104) 0.035** -0.276*** -0.198** (0.062) -0.193** (0.062) 0.033** R Square 0.328 Adjusted R Square 0.305 N 212 Abbreviations: B, unstandardized coefficient; Beta, standardized coefficient; SE, standard error; ∆R 2f, change in R square. a Parental report of child’s diagnosed learning difficulties categorized as 1) yes and 0) no or don´t know. b The highest level of parental education categorized as 1) tertiary-level education and 0) basic or upper secondary education. c Distribution of self-reported MVPA was reflected and transformed logarithmically and reflected again to restore the original order of the variable (-ln((max+1)-y)). d To measure quadratic associations between self-reported physical activity and grade point average, quadratic terms were formed with the equation: Quadratic self-reported MVPA = ((x-mean(x)) * (xmean(x)), where x was logarithmically transformed self-reported MVPA (-ln((max+1)-y)). e Distribution of self-reported screen time was transformed logarithmically (ln(y)). f The change in R square, Self-reported MVPA, Quadratic self-reported MVPA and Screen time were added to the model one by one. The level of statistical significance: *** = p<0.001, ** = p<0.01, * = p<0.05. 47 4.2 Study II: Internal consistency and stability of the CANTAB neuropsychological test battery in children Purpose of the study. Study II was conducted to ensure the internal consistency and stability of the cognitive tests before examining the associations of physical activity, sedentary behaviour and cognitive functions. The main purpose of Study II was to evaluate the internal consistency and the one-year stability of seven CANTAB tests measuring visual memory, executive function, and attention in elementary school-aged children. Methods. For this purpose, children participated in cognitive measurements in spring 2011 and again in spring 2012. The CANTAB tests Pattern Recognition Memory (PRM), Spatial Recognition Memory (SRM), Spatial Span (SSP), Stockings of Cambridge (SOC), Intra-Extra Dimensional Set Shift (IED), Reaction Time (RTI) and Rapid Visual Information Processing (RVP) were chosen according to previous studies and hypotheses about the association of physical activity and cognitive functions. In addition, a goal was to test different dimensions of cognitive functions. The internal consistency of the tests was estimated with Cronbach’s alpha reliability coefficient (α), while the stability of these tests was examined using structural equation modelling. Results. In terms of internal consistency, Cronbach’s alpha reliability coefficients were 0.65 for the PRM number of correct responses, 0.21 for the SRM number of correct responses, 0.87 for the RTI five-choice movement time, 0.66 for the RTI five-choice reaction time and 0.49 for the RVP A’. For hampering tests (SSP, SOC and IED), Cronbach’s alpha could not be determined. Stability of visual memory tests. The estimation results of the stability model for PRM are presented in Figure 3. The goodness-of-fit statistics of the constrained model for the PRM number of correct responses were good (χ2 (12) = 17.30, p = 0.14, CFI = 0.96, TLI = 0.95, RMSEA = 0.04, SRMR = 0.13). The stability for the PRM was 0.80. A stability model for SRM could not be determined, because none of the parcels loaded significantly on the hypothesized factor in the cross-sectional measurement models in 2011 and 2012. The correlation for the SRM number of correct responses between 2011 and 2012 was r(72) = 0.30 (p = 0.009). Stability of executive function tests. The stability model for the SSP could not be determined, because of the nature of the test. The correlation between the SSP span length in 2011 and in 2012 was r(72) = 0.37 (p = 0.001). The cross-sectional IRT models for SOC were estimated, and only 5 of the 12 items loaded significantly on the hypothesized factor in 2011 and only 3 of the 12 items in 2012. According to the results of the estimated baseline stability model, there was no significant association between the latent variables measured in 2011 and 2012. The correlation for SOC problems solved in the minimum number of moves between 2011 and 2012 was r(72) = 0.23 (p = 0.046). Because the latent variables measured by the errors in stages from 3 to 7 and from 8 to 9 of the IED test did not correlate with each other either in 2011 or in 2012, and the regression coefficient between the latent variables measured in consecutive years was significant only for the latter factor (b = 0.68, SE = 0.08, p < 0.001), the stages from 3 to 7 were discarded from further analyses. The tetrachoric correlation for stage 8 between 2011 and 2012 was 0.38 (p = 0.047). For stage 9, it was 0.64 (p < 0.001). 48 Stability of attention tests. The estimation results of the stability model for RTI reaction time and movement time are presented in Figure 4. The goodness-offit statistics of the constrained model for the RTI reaction time and the movement time were good (χ2 (58) = 78.51 (58) p = 0.04, CFI = 0.96, TLI = 0.95, RMSEA = 0.04, SRMR = 0.15). The stability coefficient for the RTI movement time was 0.67. For the RTI reaction time, it was 0.78. The estimation results of the stability model for RVP A’ is presented in Figure 5. The goodness-of-fit statistics of the constrained model for RVP A’ were reasonably good (χ2 (10) = 17.94 (10) p = 0.06, CFI = 0.91, TLI = 0.87, RMSEA = 0.06, SRMR = 0.25. The stability coefficient for RVP A’ was 0.62. Conclusions. Internal consistency was acceptable only in the RTI task. Oneyear stability was moderate-to-good for the PRM, RTI, and RVP. The SSP and IED showed a moderate correlation between the two measurement points. The SRM and SOC tasks were not reliable or did not provide stable measurements in this study population when comparing two measurements that were conducted one year apart. For research purposes, when using these tests, structural equation modelling is recommended to improve reliability. The results suggest that the reliability and stability of computer-based test batteries should be confirmed in the target population before using them for clinical or research purposes. FIGURE 3. Estimation results of the stability model of the Pattern Recognition Memory (PRM) test for 2011 and 2012. Standardized parameter estimates and standard errors are presented. Three parcels from 24 patterns (incorrect/correct response) of the PRM test were constructed and used as indicator variables in the confirmatory factor analyses (PRM 1, PRM 2, PRM 3). All the factor loadings were constrained to be equal to one. 49 FIGURE 4. Stability model for the Reaction Time (RTI) test for 2011 and 2012. Standardized parameter estimates and standard errors are presented. For structural equation modelling, three sub-categories of the reaction time (Reac 1, Reac 2, Reac 3) and movement time (Move 1, Move 2, Move 3) at both measurement points were formed from 15 patterns of RTI. All the factor loadings were constrained to be equal to one. 50 FIGURE 5. Stability model for the Rapid Visual Information Processing (RVP) test for 2011 and 2012. Standardized parameter estimates and standard errors are presented. The RVP (multiplied by 10) of the three blocks of the test (RVP 1, RVP 2, RVP 3) were used as indicator variables in structural equation modelling. Factor loadings were constrained to be equal across time points. 51 4.3 Study III: The associations of objectively measured physical activity and sedentary time with cognitive functions in school-aged children Purpose of the study. The main purpose of the Study III was to examine how objectively measured and self-reported physical activity and sedentary behaviour are associated with performance in cognitive tests, measuring visual memory, executive functions and attention in school-aged children. It was hypothesized that physical activity is positively, and sedentary behaviour inversely, associated with cognitive functions. Methods. On the basis of the results of Study II, which reported the reliability and stability of the CANTAB tests, the following five cognitive tests were chosen for this study: Pattern Recognition Memory (PRM), Spatial Span (SSP), Intra-Extra Dimensional Set Shift (IED), Reaction Time (RTI) and Rapid Visual Information Processing (RVP). The cross-sectional associations of self-reported physical activity and screen time and objectively measured physical activity and sedentary time with cognitive functions were examined with structural equation modelling. Results. A high level of objectively measured MVPA was associated with good performance in the RTI test (Table 6). A high level of objectively measured sedentary time was associated with good performance in the RVP test (Table 6). Objectively measured MVPA and sedentary time were not associated with other measures of cognitive functions. A high amount of self-reported computer or video game play was associated with weaker performance in the SSP test (Table 6), whereas a high amount of computer use was associated with weaker performance in the IED test (Table 7). Self-reported physical activity, total screen time, and television viewing were not associated with any measures of cognitive functions. Conclusions. In this study, objectively measured physical activity and sedentary time were positively associated with attentional processes, but not with other measured domains of cognitive functions. Self-reported computer and video game playing was negatively associated with working memory capacity, whereas computer use was negatively associated with shifting and flexibility of attention. Selfreported physical activity and total screen time were not associated with any of the cognitive tests used to measure visual memory, executive functions or attention in children. The results of the present study suggest that physical activity may benefit attentional processes, but excessive video game playing and computer use may have unfavourable effects on cognitive functions. 52 TABLE 6. Associations between children’s cognitive processes and objectively measured physical activity, sedentary time and self-reported screen time. Explanatory variables Objectively measured MVPAa Objectively measured sedentary timeb Objectively measured MVPA Objectively measured sedentary time Self-reported viewing of TV Self-reported playing of computer/video game Self-reported use of computer (other than playing) Cognitive test SE 95% CI Pf Reaction Time (RTI)c -0.130 0.062 -0.253, -0.008 0.037 -0.041 0.074 -0.186, 0.104 0.581 Rapid Visual Information Processing (RVP)d -0.040 0.093 -0.223, 0.143 0.669 0.305 0.078 0.153, 0.457 0.000 e Spatial Span (SSP) -0.003 0.067 -0.134, 0.129 0.970 -0.179 0.079 -0.333, -0.024 0.023 B 0.094 0.068 -0.040, 0.227 0.171 Abbreviations: B, estimate; SE, standard error; CI, confidence interval; p, p-value; MVPA, moderate to vigorous physical activity. a RTI measures children’s reaction times and speed of response to a visual target in milliseconds, where faster time indicates better performance. b MVPA measured with the ActiGraph accelerometer using a cut-off value of 2,296 counts per minute and expressed as min/day. c Sedentary time measured by the ActiGraph accelerometer using a cut-off value of 100 counts per minute and expressed as percentage of daily monitoring time (%/day). d RVP measures sustained attention. The score of this task is RVP A’, where range is 0.00 to 1.00; bad to good. e SSP measures length of working memory span. The score of this task is the maximum number of items that the child can successfully remember. f P-values for parameter estimates. Models have been adjusted by gender (female), the highest level of parental education (tertiary) and the child’s need for remedial education. 53 TABLE 7. Associations between children’s working memory capacity and selfreported screen time. Explanatory variables Self-reported viewing of TV (h/day)b Self-reported playing of computer/video games (h/day)b Self-reported use of computer (other than playing) (h/day)b Cognitive test Intra-Extra Dimensional Set Shift (IED) a OR 95% CI 0.868 0.623, 1.210 1.321 0.943, 1.850 0.639 0.421, 0.972 Abbreviations: OR, odds ratio; CI, confidence interval. a IED measures sifting and flexibility of attention and was categorized as 1) children who completed the test and 0) children who did not complete the test. b Self-reported viewing of television, playing of computer/video game and use of computer (other than playing) are treated as continuous variables and expressed as h/day. Model has been adjusted by gender (female), the highest level of parental education (tertiary) and the child’s need for remedial education. 54 5 DISCUSSION In the present study, the associations of physical activity and sedentary behaviour with academic achievement and cognitive functions in school-aged children were examined. Self-reported physical activity an inverse U-shaped curvilinear relationship and total screen time had an inverse linear negative relationship with academic achievement (teacher-rated GPA), whereas objectively measured physical activity or sedentary time had no association with academic achievement. High levels of objectively measured physical activity and sedentary time were associated with good attentional performance, but not with other domains of cognitive functions. Self-reported physical activity or total screen time had no association with assessed cognitive functioning. High levels of self-reported computer/video game play was associated with low visuospatial working memory performance, whereas high levels of computer use was associated with low shifting and flexibility of attention performance. Self-reported TV viewing had no association with any domains of cognitive functions. In addition, the internal consistency and stability of the seven CANTAB tests (PRM, SRM, SSP, SOC, IED, RTI, RVP) measuring visual memory, executive functions and attention were examined. Internal consistency was acceptable only in the RTI test. One-year stability was moderate-to-good for PRM, SSP, IED, RTI and RVP. The SRM and SOC were not reliable or did not provide stable measurements in the present study sample of 12-year-old healthy children. 5.1 Physical activity, academic achievement and cognitive functions The results of the present study are in line with previous studies reporting that selfreported physical activity is associated with high levels of academic performance (Stevens et al. 2008, Fox et al. 2010, Kantomaa et al. 2010, Kantomaa et al. 2013). In addition, a few previous intervention studies have reported physical activity benefitting academic performance (Donnelly et al. 2009, Davis et al. 2011, Ardoy et al. 2014). However, in the present study the relationship between self-reported MVPA and academic achievement was curvilinear. It seems that 5–6 times/week may be the optimal amount of MVPA for academic achievement. It is possible that some of the most active children spend time engaged in physical activities at the expense of time devoted to homework. According to the results of the present study, objectively measured MVPA was not associated with academic achievement supporting the study by LeBlanc et al. (2012). In contrast, Kwak et al. (2009) reported that objectively measured vigorous physical activity was positively associated with academic achievement in 16-yearsold girls, but not boys. In addition, according to Booth et al. (2014), objectively measured MVPA at the age of 11 predicted increased mother tongue performance at the age of 11, 13 and 16 for both girls and boys, and increased science performance at the age of 11 and 16 for girls. The differences between these studies may be due to slight differences in the age of the study populations and the collection and reduction decisions of objective MVPA data. In addition, in these studies, if children had 55 valid data over a minimum of 2–3 days, they were included in the analysis. Physical activity varies depending on days, whereupon a measurement period of 2–3 valid days may not be long enough to fully capture habitual physical activity. This may partly explain differing results between studies concerning the association with academic achievement. Furthermore, even though Kwak et al. (2009) and Booth et al. (2014) observed significant associations, the effect sizes and β coefficients were quite modest. Booth et al. (2014), however, speculated that results must be interpreted in context. Thus, in the Western world, where the levels of MVPA are quite low, increasing MVPA to reach an average of 60 minutes per day may result in bigger increases in academic achievement (Booth et al. 2014). Reflecting the discussion above to the results of self-reported physical activity, it seems that physical activity is an important factor affecting academic achievement, but it is not the whole story. There are also other factors affecting academic achievement, and the association between physical activity and academic achievement. According to the present results, a high level of objectively measured physical activity was associated with better performance in attention test measuring reaction time and speed of response to unpredictable stimulus. Previous studies have suggested that physical activity enhances especially inhibitory control of attention (Castelli et al. 2011, Davis et al. 2011, Fisher et al. 2011, Chaddock-Heyman et al. 2013, Spitzer & Hollmann 2013, Crova et al. 2014). However, Puder et al. (2011) reported that physical activity intervention had no effect on children’s selective attention. In addition, in the studies of Davis et al. (2011) and Fisher et al. (2011) physical activity had no effect on performance in attention scale of the Cognitive Assessment System, even though physical activity enhanced attentional inhibition performance in the antisaccade task and the Attention Network Test, respectively. Despite the earlier studies have shown that physical activity enhances inhibitory control of attention, our results suggest that physical activity may also affect lower-order levels of attention. The present results also showed that neither objectively measured nor selfreported physical activity was associated with visual memory, spatial working memory, shifting and flexibility of attention or sustained attention performance. Earlier studies have also reported that physical activity is not necessary associated with all domains of cognitive functioning, but may selectively affect executive functions and especially inhibitory control (Davis et al. 2011, Drollette et al. 2012). However, the study designs with varying age of the children as well as, the definitions, patterns and measurements of cognitive functions and physical activity, which have varied across different studies, potentially contribute to the divergent results. In the present study, the lack of association between physical activity and performance in cognitive tests may be partly due to the observation that some of the cognitive tests were not able to differentiate healthy 12-year-old children. Especially in the tests of visual memory (PRM), shifting and flexibility of attention (IED) and sustained attention (RVP), on average, children achieved very good performance. Neuropsychological test batteries were originally developed to detect largescale neurocognitive deficits, and thus, the selected tests may have been too easy to detect subtle differences in health behaviours in healthy children. Stroth et al. (2009) speculated that one reason behind their result of aerobic fitness not being associated with children’s performance in cognitive control tasks was the ceiling effect in the selected cognitive tests. 56 5.1.1 Possible mechanisms explaining the associations between physical activity, academic achievement and cognitive functions The positive association of physical activity with academic achievement and cognitive functions may be mediated by a few possible mechanisms. Physical activity may induce functional (Davis et al. 2011, Chaddock-Heyman et al. 2013) and structural (Chaddock et al. 2010a, Chaddock et al. 2010b) changes in the brain areas subserving executive functions and memory, such as hippocampus and the basal ganglia, thereby enhancing especially executive functions and memory. Physical activity has also been reported to increase levels of brain-derived neurotrophic factor (BDNF), supporting cellular processes important for learning and memory (Hopkins et al. 2012, Gomez-Pinilla & Hillman 2013). Moreover, physical activity has been shown to enhance cerebrovascular function (Brown et al. 2010). Motor function has been closely connected to children’s cognitive and academic skills and development (Iverson 2010, Davis, Pitchford & Limback 2011, Haapala et al. 2014b), and it may be an important driver in the effects of physical activity on cognitive prerequisites of learning (Kantomaa et al. 2013). In addition, some physical activities are cognitively challenging and may, therefore, facilitate cognitive function and academic performance (Best 2010, Crova et al. 2014). Obesity is another potential mediator: physical inactivity may lead to obesity, and obesity has been linked to poorer academic (Kantomaa et al. 2013) and cognitive performance (Burkhalter & Hillman 2011). Furthermore, unhealthy diet, especially overconsumption of high-caloric foods may attenuate cognitive functions via diminishing BDNF and interfering energy homeostatis (Vaynman & Gomez-Pinilla 2006, Gomez-Pinilla 2011). Participation in physical activities is often a social phenomenon offering opportunities for interaction with other children and adults, and this interaction may also have a significant impact on children’s cognitive development and learning. However, social interaction as a mediator between physical activity and cognitive or academic skills is rarely studied (Hillman, Erickson & Kramer 2008). Finally, different psychosocial factors like school contentment and self-esteem may mediated the association between physical activity and academic achievement (Kristjansson et al. 2009, Kristjansson, Sigfusdottir & Allegrante 2010). 5.1.2 Differences between objectively measured and self-reported physical activity in terms of their association with academic achievement and cognitive functions In the present study, the associations of self-reported and objectively measured physical activity with academic and cognitive performance were inconsistent. These inconsistencies may be due to differences between subjective and objective measurements of physical activity. Self-reported questionnaires have been accepted as capable of measuring physical activity (Shephard 2014). However, especially when it comes to children, it may be difficult to estimate one’s overall physical activity. In the study by Corder et al. (2010), 40% of inactive children aged 10 years old overestimated their physical activity compared to the objective measurement. 57 On the other hand, accelerometers are not capable of assessing all physical activities, such as swimming and cycling or activities that require lifting any kind of load (Strath et al. 2013). In addition, skill-specific types of physical activities based on body movements are common among children, but may not be seen in activity counts. For example, skateboarding is a skill-specific physical activity requiring balance and agility, yet it hardly accumulates activity counts. Thus, accelerometermeasured MVPA mainly illustrates cardiovascular activity with increased heart rate and respiratory frequency, while self-reported MVPA may represent different constructs and contexts of physical activity. These differences may explain inconsistent results between self-reported and objectively measured MVPA in association with academic achievement and cognitive functions. 5.2 Sedentary behaviour, academic achievement and cognitive functions The present results, which show a negative association between self-reported screen time and academic achievement, are consistent with those of previous studies (Sharif & Sargent 2006, Mößle et al. 2010, Sharif, Wills & Sargent 2010). In addition, the present results also support some of the earlier studies (Dworak et al. 2007, Swing et al. 2010) by showing that excessive video game play and computer use were associated with weaker performance in cognitive tests measuring working memory and shifting and flexibility of attention. However, in the present study, TV viewing had no association with cognitive functions, contrary to earlier studies in which excessive TV viewing has been connected to attention and learning difficulties (Johnson et al. 2007, Landhuis et al. 2007, Mizuno et al. 2012). In addition, total screen time was not associated with cognition, which is in line with other previous studies reporting no such association (Weis & Cerankosky 2010, Ferguson et al. 2012). However, earlier studies have reported screen time benefitting academic performance (Borzekowski & Robinson 2005, Jackson et al. 2006, Bittman et al. 2011, Jackson et al. 2011, Haapala et al. 2014a) and cognitive functions (Dye, Green & Bavelier 2009, Spence & Feng 2010, Boot, Blakely & Simons 2011, Granic, Lobel & Engels 2013). In sum, the results of the present study both support and contradict the earlier studies, showing that associations of screen time with cognitive and academic performance are complicated and still need clarification. These inconsistencies may be partly due to the varying designs across the studies. Especially, the age of the children is crucial: the role of screen time may be different for children in different developmental stages. For example, in Haapala et al.’s (2014a) study, television viewing had no association with academic skills and high levels of computer use and video game playing were associated with better academic skills in six to eight years old children. In contrast, in our study, high levels of screen time were associated with lower levels of academic achievement in 11–12-year-old children, suggesting that the effects of screen time on academic achievement may become emphasized in older children. In addition, the amount of screen time may be an important factor: in some studies, screen time (such as video game playing) showed benefits in cognitive functions (Dye et al. 2009). In the present study, some children spent excessive 58 amounts of time in front of screens in their free time: one fifth of the children reported having screen time of about five or more hours per day. Excessive amounts of screen time might partly explain the results suggesting that screen time is harmful to academic achievement and cognition. Excessive screen time may attenuate academic performance and cognitive functions through different mediators. Shin (2004) reported that television viewing hindered children’s academic achievement through three different pathways: time displacement, effort-passivity and attention-arousal. The time displacement theory suggests that time spent in front of a screen takes time away from intellectually demanding activities, such as doing homework and reading books, which may independently affect academic performance. Screen time is more attractive to children than school-related activities displacing comparable activities that involve learning opportunities. The effort-passivity theory hypothesizes that television watching does not require mental effort, which leads to mental laziness and passivity. Thus, children are more likely to find their way in front of television programmes that are entertaining and easy to understand, rather than engaging in activities that require intellectual functions, such as reading, and that this eventually leads to a decrease in academic achievement. The attention-arousal theory proposes that exposure to television programmes leads to superficial intellectual processing and impulsive behaviours, increasing attention difficulties and hindering academic achievement. In sum, Shin (2004) suggests that excessive amounts of screen time may displace activities involving learning opportunities and increase children’s impulsive behaviour, with the result of eventually decreasing academic skills. In addition, according to Sharif, Wills and Sargent (2010), screen exposure has an indirect effect on academic performance through increased sensation-seeking. The sensation-seeking theory touching on the time displacement and attentionarousal theories propose that there may be certain dispositions in media use, especially intense and exciting sensations, which increase desire for these kinds of experiences and are incompatible with concentrated efforts, such as reading and writing. Furthermore, attention difficulties, frequent failure to do homework, negative attitudes toward school, substance use and behaviour problems have been reported to mediate the association between television viewing and academic performance (Johnson et al. 2007, Sharif, Wills & Sargent 2010). The disadvantages and benefits of screen-based sedentary behaviour on cognitive functions and academic achievement may be explained by content of screen time, as not all types of screen time have an equal role in benefitting or impairing children’s cognitive skills and learning (Kirkorian, Wartella & Anderson 2008, Schmidt & Vandewater 2008). This might also explain the divergent results of the studies. According to Ennemoser and Schneider (2007), educational programme viewing was positively correlated with reading speed and comprehension in children, yet entertainment programme viewing was negatively correlated. Feng, Spence and Pratt (2011) reported that action-game training in young adults improved performance and attenuated gender differences favouring males in spatial attention tests, while non-action-game playing among control subjects had no effect on attention performance. Kühn et al. (2013) reported that video game playing may induce structural brain plasticity in the areas important to spatial navigation, strategic planning, 59 working memory and motor performance, which may explain the positive association between screen time and cognition. On the other hand, according to Drowak et al. (2007), interactive computer game play, but not viewing exciting films, resulted in significant declines in verbal memory performance and slow wave sleep, which is important for memory consolidation. Finally, it should be kept in mind that computer-based cognitive assessments may require similar cognitive skills as video and computer games, which would favour children who play a lot of video and computer games. The skills will develop, which are practised. To our knowledge, this was the first study that examined associations of objectively measured sedentary time with academic performance and cognitive functions. In the present study, objectively measured sedentary time was not associated with academic achievement, but had a positive association with performance in sustained attention test: children who spent more time being sedentary had higher scores in sustained attention test. This might be due to the objective measurement of sedentary time, which is a summary measure of all kinds of sedentary behaviour – including a range of various activities, such as screen time, reading, homework, interaction with friends, et cetera – but is unable to differentiate between types of sedentary behaviour. It is reasonable to suggest that some of these sedentary activities (e.g. homework and reading) benefit learning, cognition and academic achievement. 5.3 Methodological considerations 5.3.1 General strengths and limitations To our knowledge, this was the first study to examine the associations of both objectively measured and self-reported overall physical activity and sedentary behaviour on teacher-rated academic achievement and cognitive functions. From the perspective of physical activity, our study sample was quite large and representative, and showed comparable levels of physical activity in Finnish school-aged children to those reported in international results (Currie et al. 2012). The study design was cross-sectional and, therefore, conclusions regarding the causality of the observed associations cannot be drawn. Furthermore, pubertal timing, motor skills and fitness were not assessed in this study, which limits the interpretation of the results, especially concerning cognitive functions. This was also the first study to examine the internal consistency and one-year stability of CANTAB tests in healthy 12-year-old children, and thus it provided valuable and important information on the psychometric characteristics of CANTAB tests in a child population. The number of children in the follow-up measurements in 2012 was limited, which attenuated statistical power. However, there were no significant differences in the CANTAB tests according to socioeconomic status or performance between children with complete data and children who did not participate in the follow-up measurements. In addition, the models were estimated with FIML estimation, which uses all information available and takes missing data into account. From the perspective of psychometric characteristics, the age-range of the children studied was narrow limiting the generalization of the results to different 60 age groups or developmental stages. Furthermore, the study sample was culturally homogeneous including only Finnish children. It was also not possible to calculate intra-class correlations for the CANTAB tests, because the performance of the children was not measured again immediately after the first measurement. Several measurement points during the year would have given more accurate information on the effects of practice on performance in the tests. 5.3.2 Measurement of physical activity and sedentary behaviour The questions used in the present study to assess self-reported physical activity and sedentary behaviour were taken from the WHO HBSC study (Currie et al. 2012). These questions have been commonly used. Test-retest agreement has been very good for MVPA (Booth et al. 2001, Liu et al. 2010), substantial for TV viewing and computer and video game playing (Liu et al. 2010), and moderate for computer use (Liu et al. 2010). One of the main weaknesses of self-reports is accuracy due to recollection and social desirability biases. For example, according to Corder et al. (2010), children overestimated their physical activity levels compared to objectively measured physical activity. In addition, in the present study, self-reported physical activity was assessed with only one question. The question was worded so that it assesses overall MVPA during the day. However, children may not include brief spurts or incidental physical activity in overall MVPA when answering this question. Furthermore, the detailed content of screen-based sedentary behaviour was not assessed, which limits the interpretation of the results regarding screen-based sedentary behaviour. Moreover, time spent doing homework, reading, or performing other activities that may benefit academic achievement and cognition was not investigated, limiting a more precise examination of total sedentary behaviour. Children’s physical activity and sedentary time were also measured objectively with an accelerometer, and accelerometers have been considered to provide more accurate measurement than self-reports (Evenson et al. 2006). However, accelerometers do not measure all kinds of activities, such as swimming, cycling, or similar activities (Strath et al. 2013). Specifically, accelerometers that are worn on the hip, do not capture activities where load-lifting is performed (Strath et al. 2013), and do not differentiate between sitting and standing (Skotte et al. 2014). Furthermore, in the present study, children wore an accelerometer for seven days, which may not have been long enough to capture their normal physical activity. 5.3.3 Measurement of academic achievement In this study, teacher-rated academic achievement scores were used to assess children’s academic performance. Grading given by teachers enables more comprehensive evaluations of children’s academic performance, compared to grading that is based solely on standardized tests. At the same time, however, grading by teachers may result in discrepancies in grades between students (from different classes or schools) with the same level of competence. Therefore, a comparison of children’s academic achievements is not accurate if the objectives of the national curriculum 61 for each grade are not carefully followed and teachers let their biases affect grading. (Ouakrim-Soivio 2013). 5.3.4 Measurement of cognitive functions – Reliability and stability of cognitive test battery (CANTAB) Neurocognitive performance of children was assessed with the Cambridge Neuropsychological Test Automated Battery (CANTAB). There are a lot of advantages in computerized testing: computer technology has increased the efficiency, ease, and standardization of administration and saved time and money related to testing. Electronic data capture and automatic results scoring have minimized human errors in scoring and data entry and increased the accuracy of timing and response latencies. (Cernich et al. 2007, Parsey & Schmitter-Edgecombe 2013). Other advantages of computerized technology – particularly for the measurement of attention, motor, and memory functioning – are the availability of almost unlimited alternate forms and an increased number of trials. This minimizes practice effects and allows more assessments at shorter time intervals compared to traditional measures. A large number of trials and accurate assessment of reaction times results in data that is normally distributed and on a true interval scale (Betts et al. 2006). This is particularly important when slight changes in performance across time are assessed. Computer technology also provides test administration conditions that are accommodating for individuals with particular needs (American Educational Research Association et al. 2014). Touch-screen technology facilitates use by young children and certain clinical groups, and allows more reliable assessment of motor function and processing speed compared to traditional measures using an individual administrator wielding a stopwatch. In addition, non-verbal culture-neutral test stimuli are often used, which allow the application of computerized test batteries (like CANTAB) for individuals from different racial, ethnic, geographic, or sociocultural backgrounds. (Luciana 2003, Henry 2010). CANTAB also has simple standardized test administration; thus, it is easy to use and no previous IT or scientific training is needed to set-up and administer the test (Cambridge Cognition Ltd. 2006). However, there is limited information about the psychometric properties of CANTAB and other computerized batteries. This study was one of only a few studies to measure the psychometric properties of CANTAB tests, especially in children. Our results, which show that the internal consistency of certain CANTAB tests (PRM, SRM, RVP) was quite low and acceptable only for the RTI test, do not support the study by Luciana et al. (2003), who reported that the internal consistency of CANTAB tests has been uniformly high in 4–12-yearold children. Internal consistency could not be determined for IED or SSP because of the hampering nature of these tests. The results of the present study are in line with previous studies measuring the test-retest stability of CANTAB tests in both children (Gau & Shang 2010, Fisher et al. 2011) and adults (Lowe & Rabbitt 1998, Henry & Bettenay 2010). In this study, one-year stability was moderate in certain tests (PRM, SSP, IED, RTI, RVP). However, earlier studies measured test-retest correlation with a time-interval of a few weeks, whereas in the present study the time-interval was approximately one year. In the present study, SRM and SOC were not reliable or stable measures in healthy 12-year-old children over one year period. 62 Low internal consistency might be a result of the ceiling effect, especially in the case of PRM and RVP. In these tests, children have been reported to reach ceiling levels by the age of 7 and after 10 years of age, respectively (Halperin et al. 1991, Luciana & Nelson 2002). In addition, in the present study, 12-year-old healthy children reached ceiling levels also in the IED test, which is in line with earlier studies (Luciana & Nelson 1998, Anderson 2002, Luciana & Nelson 2002). Thus, when ceiling levels are reached, the errors the children do may be sparse and random, which may explain why the patterns of the tests do not have a high correlation with each other. It may be problematic to discriminate children with high ability and internal consistency may be affected. Stability could also be affected by the ceiling effect and a long test-retest interval. In this study, stability was measured with a one-year interval. It is to be expected that the cognitive abilities of 11-year-old children develop over the course of a year. Therefore, measuring their performance twice – in the beginning with a shorter interval between the measurements – would have enabled calculation of test-retest reliability, thus ruling out developmental effects. The SOC test, which measures spatial planning and spatial working memory, and is identical to the traditional Tower of London task, did not measure the phenomena it was supposed to with satisfying reliability. Previous studies have also raised questions about the psychometric characteristics of both traditional and CANTAB versions of this test: previous studies have reported low internal consistency (Humes et al. 1997, Ahonniska et al. 2000) and test-retest reliability (Lowe & Rabbitt 1998, Bishop et al. 2001). However, temporal stability has also been reported to be acceptable (Gnys & Willis 1991, Ahonniska et al. 2000, Gau & Shang 2010). Both large intra-individual variation due to different rates of learning and task novelty have been suggested as explanations for low reliability and stability (Lowe & Rabbitt 1998, Ahonniska et al. 2000). Children’s performance in the SOC test improves approximately at the age of 11 years due to developmental spurt (Anderson, Anderson & Lajoie 1996). In addition, performance in the tests of executive function can abruptly improve when an individual discovers an optimal strategy, but there is less or no chance in performance if a strategy is not found, and it may even decline if an incorrect strategy is attempted (Lowe & Rabbitt 1998). In the present study, 58% showed improved performance, and 31% showed a decline in performance. These different practice effects may weaken test-retest reliability (Lowe & Rabbitt 1998). CANTAB tests as well as other computer-based tests are based on traditional neuropsychological tests. However, a few studies have shown that the computerized versions of the tests are not equivalent to traditional ones (Feldstein et al. 1999, Smith et al. 2013). The reason for these findings is unknown, and it can only be speculated that computerized test sessions are more prone to attentional disruptions than manual sessions, or they may not offer similar perceptual characteristics as manual versions. Despite the potential advantages that computerized technology offers for neuropsychological testing – especially the ease of building ready algorithms for calculating indexes and scores – computer-assisted assessment also produces a risk factor. Due to commercial competitive reasons, the companies providing these tools may not always publish detailed information about these algorithms or the psychometric properties of the tasks. The general characteristics of scoring algorithms and 63 the accuracy of the algorithms as well as technical evidence should be documented and reviewed periodically (American Educational Research Association et al. 2014). According to the Finnish Psychological Test Committee (2013), test batteries (computerized or not) that do not provide a satisfactory level of information about the psychometric properties of the tasks in the technical manual should not be used in clinical practice. Likewise, in clinical practice, an expert (a psychologist or an MD) should always do the interpretation of the test results. There are two reasons for this. First, to use CANTAB or other computerized test batteries, there needs to be a person who supervises and paces the introduction of tests and monitors the assessment process (Luciana 2003, American Educational Research Association et al. 2014). Secondly, the interpretation and the clinical conclusions and decisions made based partly on the test results produce a juridical situation where there needs to be a responsible decision-maker. Most countries do not allow automated decisionmaking in healthcare. There will be rapid growth in the usage of computer-assisted test batteries. However, little is known about the psychometric properties of computerized batteries as well as how performance on computerized batteries correlates with traditional neuropsychological measures. In addition, it is also vital to have more information about the reliability and validity of these batteries in all clinical target populations of interest as well as in samples derived from the normal population. 64 6 CONCLUSIONS 6.1 Summary and main conclusions The main findings and conclusions of each aim of the study can be summarized as follows: 1. Self-reported (but not objectively measured) physical activity and screenbased sedentary behaviour were associated with academic achievement in school-aged children. Self-reported MVPA had an inverse curvilinear relationship with the grade point averages of school subjects, while self-reported screen time had an inverse linear relationship with grade point averages. Objective and subjective measurements may reflect different constructs and contexts of physical activity and sedentary behaviour in association with academic outcomes. 2. The internal consistency of most of the seven CANTAB tests used in this study was below the accepted level of 0.7. The one-year stability of most tests was moderate-to-good. In addition, in the present study population of 12-yearold healthy children, the ceiling levels were reached in the PRM, IED, and RVP tests. The psychometric characteristics of traditional neuropsychological tests may be lost when converted into computer form; this should be confirmed among target populations before using computer-based test batteries for clinical or research purposes. The results of the present study suggest that the use of structural equation modelling would improve the reliability of these tests when using them in research analysis. 3. High objectively measured physical activity and sedentary time were associated with good performance in the test measuring attentional processes, but not in the tests related to other domains of cognitive functioning. Self-reported physical activity, total screen time or television viewing were not associated with any of the cognitive tests measuring visual memory, executive functions or attention in children. High self-reported time spent in computer or video game play and computer use was associated with poor performance in the tests measuring visuospatial working memory and shifting and flexibility of attention, respectively. The results of the present study suggest that physical activity may benefit some attentional processes. However, excessive video game play and computer use may have an unfavourable influence on some cognitive functions. However, all sedentary time is not harmful to cognition, but may include activities that benefit certain cognitive functions. 65 6.2 Implications and future directions This doctoral thesis provide important information about the associations of physical activity and sedentary life with cognitive functions and academic performance in Finnish school-aged children, in order for both practitioners and policy-makers to better develop learning, education and health in the Finnish school system. The topic is highly relevant and currently important. The results suggest that in Finland, which has an educational system of high quality and equality, physical activity may enhance some attentional processes and academic achievement, whereas excessive screen-based sedentary behaviour may attenuate them. As physical inactivity among children is increasing worldwide, physical activity that benefits both health and learning is becoming an important part of education. The enhancement of physical activity in school settings has recently also been the focus of Finnish Sport policy aimed at improving the health and well-being of young people. Almost every Finnish child conducts their compulsory education in school, spending plenty of time there during the week. Thus, schools play an important role in encouraging children to be more physically active, both at school and outside school hours. As learning is the centre of action at school, research that shows a positive association between physical activity and learning may also pique the interest of parents and teachers who are not into physical activity themselves. A few studies have also shown that both acute exercise and chronic physical activity enhances cognitive performance, especially in children with ADHD (Verret et al. 2012, Pontifex et al. 2013) and in children with weaker cognitive performance (Reynolds & Nicolson 2007, Drollette et al. 2014). This important observation may significantly contribute to children’s learning and well-being in schools. Providing a learning environment where physical activity is implemented in alignment with learning objectives, may establish new opportunities to support learning and individual development. There are many options when integrating physical activity in the school day. However, it is important that physical activity is implemented in a manner that supports adoption of a physically active lifestyle. This is especially important, because physical activity in childhood and adolescence predicts physical activity in adulthood (Telama et al. 2014). Creating positive physical activity patterns that offer good experiences and support autonomy, competence and relatedness in childhood enhances physical activity over the course of one’s entire life (Hirvensalo & Lintunen 2011). Promoting a physically active lifestyle may also support lifelong academic and cognitive performance and learning. Although, excessive screen time may have unfavourable effect on academic performance and cognition, a reasonable amount of video game play and computer use may benefit them (Bittman et al. 2011, Boot, Blakely & Simons 2011, Jackson et al. 2011, Granic, Lobel & Engels 2013). Today, exergames that combine physical activity and video games are popular, and they are considered to provide one possible way to decrease children’s excessive sedentary behaviour (Staiano & Calvert 2011). In addition, exergaming may also enhance cognitive functions, especially executive functions compared to non-players (Staiano, Abraham & Calvert 2012). These kinds 66 of approaches, which combine physical activity and media use, may make important contributions to future learning environments. The advance of comprehensive approaches may support children’s physical, psychological and social growth, as well as development and learning. Although research examining the association of physical activity and sedentary behaviour with academic and cognitive performance has increased during recent years, the studies, including the current thesis, have mainly been cross-sectional. Thus, studies with longitudinal study designs and randomized controlled trials are needed to clarify the causal relationship of these variables. In addition, it would be valuable in future studies to specify the mechanisms mediating and underlying these associations. The few studies to date concerning these mechanisms have largely focused on physiological and neural mechanisms like cerebrovascular function, levels of growth factors and neural activity in the brain, which are very important factors for explaining the relationship between physical activity and cognition and academic performance. However, in the future, social interaction and context-related factors should also be considered and taken into account. The mechanisms and mediators behind the association between sedentary behaviour and cognitive and academic performance may be even more complicated and in need of clarification. Furthermore, more information is needed not only on the amount but also the types and contexts of physical activity and sedentary behaviour, which affect academic performance and specific kinds of cognition. It would be important to examine the dose-response relationship of both physical activity and screen-based sedentary behaviour with academic achievement and cognitive functions. This study provides important insight into inconsistencies between self-reported and objective measurements of physical activity and sedentary time in association with academic and cognitive performance. The results of the present study recommend that future studies use both subjective and objective measurements to examine physical activity, sedentary behaviour, and academic achievement. In addition, a longer period of wearing the accelerometer and, preferably, several measurement points during the school year should be considered. Only a few studies have measured a broad range of cognitive functions. The associations between physical activity and different dimensions of cognitive functions are somewhat inconsistent, highlighting the need for new studies to clarify the benefits of physical activity on a wide range of cognitive performance. In addition, the definitions, patterns and measurements of cognitive functions have varied across different studies, which makes it difficult to directly compare and summarize the results of these earlier studies and to determine the specific benefits and disadvantages that physical activity and sedentary behaviour may have on cognitive functions. The use of computerized test batteries to measure cognitive functions is increasing worldwide. As there are hardly any studies that assess the internal consistency and stability of these test batteries in a child population, and the results of the present study suggest that the confirmation of psychometric properties of these test batteries in the target population is highly important, more research is needed 67 to assess the reliability and stability of computerized test batteries. In particular, research on different age groups and developmental stages is needed. Technology, which is developing with dizzying pace imposes new challenges also for research assessing the effects of screen-bases sedentary time. Children today are more familiar with different media devices than previous generations. In Finland, 95% of 10–12-year-old children are estimated to have access to electronic media at home (Suoninen 2013). This may have an influence on results concerning cognitive functions, especially when computerized test batteries are used. In the future, it might be reasonable to use also other possible alternatives, such as everyday real-world cognitive tasks that could differentiate children’s cognitive performance. In addition, computers, tablets, and smartphones are always available, which makes it difficult to assess how much time is spent in front of the screens. Furthermore, the dynamic nature of screen time and media use makes it difficult to predict the longitudinal effects of screen time. For example, younger and younger children use these technologies daily, so the effects of screen time may be significantly different for those two-year-old children (who develop with technology) when they are of school age, than for school-age children now. Even two-year-old children today have gained different experience from technology than two-year-old children two years ago. The history of screen time may be one important factor to take into account in future studies assessing the association between screen time and cognition. The results of this study indicate that physical activity is positively associated and screen time inversely associated with academic performance and certain cognitive functions in children, supporting the importance of promoting physical activity in school settings and a physically active lifestyle overall. In the future, research that focuses on specifying, deepening and applying this information is needed. 68 REFERENCES Ahamed, Y., MacDonald, H., Reed, K., Naylor, P., Liu-Ambrose, T. & McKay, H. 2007. School-based physical activity does not compromise children's academic performance. Medicine and Science in Sports and Exercise 39 (2), 371–376. Ahonniska, J., Ahonen, T., Aro, T., Tolvanen, A. & Lyytinen, H. 2000. Repeated assessment of the Tower of Hanoi Test: Reliability and age effects. Assessment 7 (3), 297–310. American Educational Research Association, American Psychological Association, National Council on Measurement in Education 2014. Standards for educational and psychological testing. Washington DC: American Educational Research Association. Anderson, D. R. & Collins, P. A. 1988. The impact on children's education: Television's influence on cognitive development. Working Paper No. 2, 1–98. Anderson, O. R. 1997. A neurocognitive perspective on current learning theory and science instructional strategies. Science Education 81 (1), 67–89. Anderson, P. 2002. Assessment and development of executive function (EF) during childhood. Child Neuropsychology 8 (2), 71–82. Anderson, P., Anderson, V. & Lajoie, G. 1996. The tower of London test: Validation and standardization for pediatric populatons. The Clinical Neuropsychologist 10 (1), 54–65. Ardoy, D., Fernández-Rodríguez, J., Jiménez-Pavón, D., Castillo, R., Ruiz, J. & Ortega, F. 2014. A physical education trial improves adolescents' cognitive performance and academic achievement: The EDUFIT study. Scandinavian Journal of Medicine & Science in Sports 24 (1), 52–61. Armstrong, N., Tomkinson, G. & Ekelund, U. 2011. Aerobic fitness and its relationship to sport, exercise training and habitual physical activity during youth. British Journal of Sports Medicine 45 (11), 849–858. Ansley, T. 1997. The role of standardized achievement tests in grades K-12. In G. D. Phye (Ed.), Handbook of classroom assessment: Learning, achievement, and adjustment. San Diego: Academic Press, 265–285. Baars, B. J. & Gage, N. M. 2010. Cognition, brain, and consciousness: Introduction to cognitive neuroscience. Academic Press. Baddeley, A. 2012. Working memory: theories, models, and controversies. Annual Review of Psychology 63, 1–29. Baddeley, A. D. & Hitch, G. J. 1974. Working memory. The Psychology of Learning and Motivation 8, 47–9. Bandura, A. & McClelland, D. C. 1977. Social learning theory, Englewood Cliffs, NJ: Prentice-Hall. Barnes, J., Behrens, T. K., Benden, M. E., Biddle, S., Bond, D., Brassard, P., Brown, H., Carr, L., Chaput, J. & Christian, H. 2012. Letter to the Editor: Standardized use of the terms “sedentary” and “sedentary behaviours”. Applied Physiology Nutrition and Metabolism 37 (3), 540–542. Beentjes, J. W. & Van der Voort, T.H.A. 1988. Television's impact on children's reading. Reading Research Quarterly 23 (4), 389–413. 69 Best, J. R. 2010. Effects of physical activity on children’s executive function: Contributions of experimental research on aerobic exercise. Developmental Review 30 (4), 331–351. Best, J. R. 2012. Exergaming immediately enhances children's executive function. Developmental Psychology 48 (5), 1501–1510. Best, J. R. & Miller, P. H. 2010. A developmental perspective on executive function. Child Development 81 (6), 1641–1660. Best, J. R., Miller, P. H. & Jones, L. L. 2009. Executive functions after age 5: Changes and correlates. Developmental Review 29 (3), 180–200. Betts, J., McKay, J., Maruff, P. & Anderson, V. (2006). The development of sustained attention in children: The effect of age and task load. Child Neuropsychology 12 (3), 205–221. Bishop, D., Aamodt-Leeper, G., Creswell, C., McGurk, R. & Skuse, D. 2001. Individual differences in cognitive planning on the Tower of Hanoi task: Neuropsychological maturity or measurement error? Journal of Child Psychology and Psychiatry 42 (4), 551–556. Bittman, M., Rutherford, L., Brown, J. & Unsworth, L. 2011. Digital natives? New and old media and children's outcomes. Australian Journal of Education 55 (2), 161– 175. Blackwell, A. D., Sahakian, B. J., Vesey, R., Semple, J. M., Robbins, T. W. & Hodges, J. R. 2004. Detecting dementia: Novel neuropsychological markers of preclinical Alzheimer's disease. Dementia and Geriatric Cognitive Disorders 17 (1–2), 42–48. Blom, L., Alvarez, J., Zhang, L. & Kolbo, J. 2011. Associations between health-related physical fitness, academic achievement and selected academic behaviors of elementary and middle school students in the State of Mississippi. Journal of Research 6 (1), 13–19. Boot, W. R., Blakely, D. P. & Simons, D. J. 2011. Do action video games improve perception and cognition? Frontiers in Psychology 2. Booth, J. N., Leary, S. D., Joinson, C., Ness, A. R., Tomporowski, P. D., Boyle, J. M. & Reilly, J. J. 2014. Associations between objectively measured physical activity and academic attainment in adolescents from a UK cohort. British Journal of Sports Medicine 48 (3), 265–270. Booth, M., Okely, A., Chey, T. & Bauman, A. 2001. The reliability and validity of the physical activity questions in the WHO health behaviour in schoolchildren (HBSC) survey: A population study. British Journal of Sports Medicine 35 (4), 263–267. Borzekowski, D. L. G. & Robinson, T. N. 2005. The remote, the mouse, and the no. 2 pencil: The household media environment and academic achievement among third grade students. Archives of Pediatrics & Adolescent Medicine 159 (7), 607–613. Brown, A. D., McMorris, C. A., Longman, R. S., Leigh, R., Hill, M. D., Friedenreich, C. M. & Poulin, M. J. 2010. Effects of cardiorespiratory fitness and cerebral blood flow on cognitive outcomes in older women. Neurobiology of Aging 31 (12), 2047– 2057. Buck, S. M., Hillman, C. H. & Castelli, D. M. 2008. The relation of aerobic fitness to stroop task performance in preadolescent children. Medicine and Science in Sports and Exercise 40 (1), 166–172. 70 Budde, H., Voelcker-Rehage, C., Pietraßyk-Kendziorra, S., Ribeiro, P. & Tidow, G. 2008. Acute coordinative exercise improves attentional performance in adolescents. Neuroscience Letters 441 (2), 219–223. Burkhalter, T. M. & Hillman, C. H. 2011. A narrative review of physical activity, nutrition, and obesity to cognition and scholastic performance across the human lifespan. Advances in Nutrition 2 (2), 201S–206S. Caprara, G. V., Vecchione, M., Alessandri, G., Gerbino, M. & Barbaranelli, C. 2011. The contribution of personality traits and self-efficacy beliefs to academic achievement: A longitudinal study. British Journal of Educational Psychology 81 (1), 78–96. Carlson, S. A., Fulton, J. E., Lee, S. M., Maynard, L. M., Brown, D. R., Kohl, H. W.,3 rd & Dietz, W. H. 2008. Physical education and academic achievement in elementary school: Data from the early childhood longitudinal study. American Journal of Public Health 98 (4), 721–727. Caspersen, C. J., Powell, K. E. & Christenson, G. M. 1985. Physical activity, exercise, and physical fitness: Definitions and distinctions for health-related research. Public Health Reports 100 (2), 126–131. Castelli, D. M., Hillman, C. H., Buck, S. M. & Erwin, H. E. 2007. Physical fitness and academic achievement in third-and fifth-grade students. Journal of Sport & Exercise Psychology 29 (2), 239–252. Castelli, D.M, Hillman, C.H, Hirsch, J., Hirsch, A. & Drollette, E. 2011. FIT Kids: Time in target heart zone and cognitive performance. Preventive Medicine 52, S55– S59. Catering, M. C. & Polak, E. D. 1999. Effects of two types of activity on the performance of second-, third-, and fourth-grade students on a test of concentration. Perceptual and Motor Skills 89 (1), 245–248. Centers for Disease Control and Prevention. 2010. The association between schoolbased physical activity, including physical education, and academic performance. Atlanta, GA: U.S. Department of Health and Human Services. Cernich, A. N., Brennana, D. M., Barker, L. M. & Bleiberg, J. 2007. Sources of error in computerized neuropsychological assessment. Archives of Clinical Neuropsychology 22 (Suppl 1), S39–S48. Chaddock, L., Erickson, K. I., Prakash, R. S., Kim, J. S., Voss, M. W., VanPatter, M., Pontifex, M. B., Raine, L. B., Konkel, A., Hillman, C. H., Cohen, N. J. & Kramer, A. F. 2010a. A neuroimaging investigation of the association between aerobic fitness, hippocampal volume, and memory performance in preadolescent children. Brain Research 1358, 172–183. Chaddock, L., Erickson, K. I., Prakash, R. S., VanPatter, M., Voss, M. W., Pontifex, M. B., Raine, L. B., Hillman, C. H. & Kramer, A. F. 2010b. Basal ganglia volume is associated with aerobic fitness in preadolescent children. Developmental Neuroscience 32, 249–256. Chaddock, L., Erickson, K. I., Prakash, R. S., Voss, M. W., VanPatter, M., Pontifex, M. B., Hillman, C. H. & Kramer, A. F. 2012a. A functional MRI investigation of the association between childhood aerobic fitness and neurocognitive control. Biological Psychology 89 (1), 260–268. Chaddock, L., Hillman, C. H, Buck, S. M. & Cohen, N. J. 2011. Aerobic fitness and executive control of relational memory in preadolescent children. Medicine & Science in Sports & Exercise 43 (2), 344–349. 71 Chaddock, L., Hillman, C. H., Pontifex, M. B., Johnson, C. R., Raine, L. B. & Kramer, A. F. 2012b. Childhood aerobic fitness predicts cognitive performance one year later. Journal of Sports Sciences 30 (5), 421–430. Chaddock, L., Neider, M. B., Lutz, A., Hillman, C. H. & Kramer, A. F. 2012c. Role of childhood aerobic fitness in successful street crossing. Medicine and Science in Sports and Exercise 44 (4), 749–753. Chaddock-Heyman, L., Erickson, K. I., Voss, M. W., Knecht, A. M., Pontifex, M. B., Castelli, D. M., Hillman, C. H. & Kramer, A. F. 2013. The effects of physical activity on functional MRI activation associated with cognitive control in children: a randomized controlled intervention. Frontiers in Human Neuroscience 7. Chernin, A. R. & Linebarger, D. L. 2005. The relationship between children’s television viewing and academic performance. Archives of Pediatrics & Adolescent Medicine 159 (7), 687–689. Chic, C. & Chen, J. 2011. The relationship between physical education performance, fitness tests and academic achievement in elementary school. The International Journal of Sport and Society 2 (1), 65–73. Chomitz, V., Slining, M., McGowan, R., Mitchell, S., Dawson, G. & Hacker, K. 2009. Is there a relationship between physical fitness and academic achievement? Positive results from public school children in the Northeastern United States. Journal of School Health 79 (1), 30–37. Christakis, D. A. 2009. The effects of infant media usage: What do we know and what should we learn? Acta Paediatrica 98 (1), 8–16. Christakis, D. A., Zimmerman, F. J., DiGiuseppe, D. L. & McCarty, C. A. 2004. Early television exposure and subsequent attentional problems in children. Pediatrics 113 (4), 708–713. Christopher Westland, J. 2010. Lower bounds on sample size in structural equation modeling. Electronic Commerce Research and Applications 9 (6), 476–487. Clarke, A. T. & Kurtz-Costes, B. 1997. Television viewing, educational quality of the home environment, and school readiness. The Journal of Educational Research 90 (5), 279–285. Coe, D. P., Pivarnik, J. M., Womack, C. J., Reeves, M. J. & Malina, R. M. 2006. Effect of physical education and activity levels on academic achievement in children. Medicine and Science in Sports and Exercise 38 (8), 1515–1519. Coe, D. P., Pivarnik, J. M., Womack, C. J., Reeves, M. J. & Malina, R. M. 2004. Role of physical education and activity on academic achievement in middle school children. Medicine & Science in Sports & Exercise 36 (5), S102. Cooper, H., Valentine, J. C., Nye, B. & Lindsay, J. J. 1999. Relationships between five after-school activities and academic achievement. Journal of Educational Psychology 91 (2), 369–378. Corder, K., van Sluijs, E. M. F., McMinn, A. M., Ekelund, U., Cassidy, A. & Griffin, S. J. 2010. Perception versus reality: Awareness of physical activity levels of British children. American Journal of Preventive Medicine 38 (1), 1–8. Crova, C., Struzzolino, I., Marchetti, R., Masci, I., Vannozzi, G., Forte, R. & Pesce, C. 2014. Cognitively challenging physical activity benefits executive function in overweight children. Journal of Sports Sciences 32 (3), 201–211. 72 Currie, C., Zanotti, C., Morgan, A., Currie, D., De Looze, M., Roberts, C., Samdal, O., Smith, O. R. F. & Barnekow, V. 2012. Social determinants of health and well-being among young people. Health behaviour in school-aged children (HBSC) study. International report from the 2009/2010 survey. Health policy for children and adolescents (6). Daley, A. J. & Ryan, J. 2000. Academic performance and participation in physical activity by secondary school adolescents. Perceptual and Motor Skills 91 (2), 531– 534. Davis, C. & Cooper, S. 2011. Fitness, fatness, cognition, behavior, and academic achievement among overweight children: Do cross-sectional associations correspond to exercise trial outcomes? Preventive Medicine 52, S65–S69. Davis, C. L., Tomporowski, P. D., Boyle, C. A., Waller, J. L., Miller, P. H., Naglieri, J. A. & Gregoski, M. 2007. Effects of aerobic exercise on overweight children's cognitive functioning: A randomized controlled trial. Research Quarterly for Exercise and Sport 78 (5), 510–519. Davis, C. L., Tomporowski, P. D., McDowell, J., Austin, B., Miller, P. H., Yanasak, N., Allison, J. & Naglieri, J. A. 2011. Exercise improves executive function and achievement and alters brain activation in overweight children: A randomized, controlled trial. Health Psychology 30 (1), 91–98. Davis, E. E., Pitchford, N. J. & Limback, E. 2011. The interrelation between cognitive and motor development in typically developing children aged 4–11 years is underpinned by visual processing and fine manual control. British Journal of Psychology 102 (3), 569–584. de Jong, P. F. 1999. Hierarchical regression analysis in structural equation modeling. Structural Equation Modeling: A Multidisciplinary Journal 6 (2), 198–211. Dexter, T. 1999. Relationships between sport knowledge, sport performance and academic ability: Empirical evidence from GCSE Physical Education. Journal of Sports Sciences 17 (4), 283–295. Diamond, A. 2013. Executive functions. Annual Review of Psychology 64, 135–168. Dollman, J., Boshoff, K. & Dodd, G. 2006. The relationship between curriculum time for physical education and literacy and numeracy standards in South Australian primary schools. European Physical Education Review 12 (2), 151–163. Donnelly, J., Greene, J., Gibson, C., Smith, B., Washburn, R., Sullivan, D., DuBose, K., Mayo, M., Schmelzle, K., Ryan, J., Jacobsen, D. & Williams, S. 2009. Physical Activity Across the Curriculum (PAAC): A randomized controlled trial to promote physical activity and diminish overweight and obesity in elementary school children. Preventive Medicine 49 (4), 336–341. Drollette, E. S., Scudder, M. R., Raine, L. B., Moore, R. D., Saliba, B. J., Pontifex, M. B. & Hillman, C. H. 2014. Acute exercise facilitates brain function and cognition in children who need it most: An ERP study of individual differences in inhibitory control capacity. Developmental Cognitive Neuroscience 7, 53–64. Drollette, E. S., Shishido, T., Pontifex, M. B. & Hillman, C. H. 2012. Maintenance of cognitive control during and after walking in preadolescent children. Medicine and Science in Sport and Exercise 44 (10), 2017–2024. Duncan, M. & Johnson, A. 2014. The effect of differing intensities of acute cycling on preadolescent academic achievement. European Journal of Sport Science 14 (3), 279–286. 73 Dworak, M., Schierl, T., Bruns, T. & Strüder, H. K. 2007. Impact of singular excessive computer game and television exposure on sleep patterns and memory performance of school-aged children. Pediatrics 120 (5), 978–985. Dwyer, T., Sallis, J. F., Blizzard, L., Lazarus, R. & Dean, K. 2001. Relation of academic performance to physical activity and fitness in children. Pediatric Exercise Science 13 (3), 225–237. Dye, M. W., Green, C. S. & Bavelier, D. 2009. Increasing speed of processing with action video games. Current Directions in Psychological Science 18 (6), 321–326. Edwards, J., Mauch, L. & Winkelman, M. 2011. Relationship of nutrition and physical activity behaviors and fitness measures to academic performance for sixth graders in a Midwest City school district. Journal of School Health 81 (2), 65– 73. Eitle, T. M. 2005. Do gender and race matter? Explaining the relationship between sports participation and achievement. Sociological Spectrum 25 (2), 177–195. Eitle, T. M. & Eitle, D. J. 2002. Race, cultural capital, and the educational effects of participation in sports. Sociology of Education 75, 123–146. Ekelund, U., Luan, J., Sherar, L. B., Esliger, D. W., Griew, P. & Cooper, A. 2012. Moderate to vigorous physical activity and sedentary time and cardiometabolic risk factors in children and adolescents. JAMA: The Journal of the American Medical Association 307 (7), 704–712. Ekelund, U., Tomkinson, G. & Armstrong, N. 2011. What proportion of youth are physically active? Measurement issues, levels and recent time trends. British Journal of Sports Medicine 45 (11), 859–865. Ennemoser, M. & Schneider, W. 2007. Relations of television viewing and reading: Findings from a 4-year longitudinal study. Journal of Educational Psychology 99 (2), 349–368. Erickson, K. I., Voss, M. W., Prakash, R. S., Basak, C., Szabo, A., Chaddock, L., Kim, J. S., Heo, S., Alves, H., White, S. M., Wojcicki, T. R., Mailey, E., Vieira, V. J., Martin, S. A., Pence, B. D., Woods, J. A., McAuley, E. & Kramer, A. F. 2011. Exercise training increases size of hippocampus and improves memory. Proceedings of the National Academy of Sciences of the United States of America 108 (7), 3017–3022. Ericsson, I. 2008. Motor skills, attention and academic achievements. An intervention study in school years 1–3. British Educational Research Journal 34 (3), 301–313. Espinoza, F. 2009. Using project-based data in physics to examine television viewing in relation to student performance in science. Journal of Science Education and Technology 18 (5), 458–465. Evenson, K. R., Catellier, D. J., Gill, K., Ondrak, K. S. & McMurray, R. G. 2006. Calibration of two objective measures of physical activity for children. Journal of Sports Sciences 24 (14), 1557–1565. Feldstein, S. N., Keller, F. R., Portman, R. E., Durham, R. L., Klebe, K. J. & Davis, H. P. 1999. A comparison of computerized and standard versions of the Wisconsin Card Sorting Test. The Clinical Neuropsychologist 13 (3), 303–313. Feng, J., Spence, I. & Pratt, J. 2007. Playing an action video game reduces gender differences in spatial cognition. Psychological Science 18 (10), 850–855. 74 Ferguson, C. J., Garza, A., Jerabeck, J., Ramos, R. & Galindo, M. 2012. Not worth the fuss after all? Cross-sectional and prospective data on violent video game influences on aggression, visuospatial cognition and mathematics ability in a sample of youth. Journal of Youth and Adolescence 42, 109–122. The Finnish Psychological Test Committee. 2013. Ohjeistus tietokoneavusteisten testien käyttämisestä psykologisessa tutkimuksessa [Guidelines for using computer-assisted tests in psychological assessment] (In Finnish). Available at http://www.psyli.fi/files/1445/Tietokoneavusteinen_psykologinen_tutkimus _18.12.2012.pdf Fisher, A., Boyle, J., Paton, J., Tomporowski, P., Watson, C., McColl, J. & Reilly, J. 2011. Effects of a physical education intervention on cognitive function in young children: Randomized controlled pilot study. BMC Pediatrics 11 (97). Fox, C., Barr-Anderson, D., Neumark-Sztainer, D. & Wall, M. 2010. Physical activity and sports team participation: Associations with academic outcomes in middle school and high school students. Journal of School Health 80 (1), 31–37. Fredericks, C. R., Kokot, S. J. & Krog, S. 2006. Using a developmental movement programme to enhance academic skills in grade 1 learners. South African Journal for Research in Sport, Physical Education and Recreation 28 (1), 29–42. Fried, R., Hirshfeld-Becker, D., Petty, C., Batchelder, H. & Biederman, J. 2012. How informative is the CANTAB to assess executive functioning in children with ADHD? A Controlled Study. Journal of Attention Disorders. Gaddy, G. D. 1986. Television's impact on high school achievement. Public Opinion Quarterly 50 (3), 340–359. Gallotta, M. C., Guidetti, L., Franciosi, E., Emerenziani, G. P., Bonavolontà, V. & Baldari, C. 2012. Effects of varying type of exertion on children’s attention capacity. Medicine and Science in Sports and Exercise 44, 550–555. Gau, S. S. & Shang, C. 2010. Executive functions as endophenotypes in ADHD: evidence from the Cambridge Neuropsychological Test Battery (CANTAB). Journal of Child Psychology and Psychiatry 51 (7), 838–849. Gentile, D. A. & Walsh, D. A. 2002. A normative study of family media habits. Journal of Applied Developmental Psychology 23 (2), 157–178. Gnys, J. A. & Willis, W. G. 1991. Validation of executive function tasks with young children. Developmental Neuropsychology 7 (4), 487–501. Gomez-Pinilla, F. 2011. The combined effects of exercise and foods in preventing neurological and cognitive disorders. Preventive Medicine 52, S75–S80. Gomez-Pinilla, F. & Hillman, C. 2013. The influence of exercise on cognitive abilities. Comprehensive Physiology 52, S75–S80. Gortmaker, S. L., Salter, C. A., Walker, D. K. & Dietz, W. H. 1990. The impact of television viewing on mental aptitude and achievement: A longitudinal study. Public Opinion Quarterly 54 (4), 594–604. Granic, I., Lobel, A. & Engels, R. C. 2013. The benefits of playing video games. American Psychologist 69(1), 66–78. Grissom, J. B. 2005. Physical fitness and academic achievement. Journal of Exercise Physiology Online 8 (1). Guiney, H. & Machado, L. 2013. Benefits of regular aerobic exercise for executive functioning in healthy populations. Psychonomic Bulletin & Review 20 (1), 73– 86. 75 Haapala, E. A., Poikkeus, A. M., Kukkonen-Harjula, K., Tompuri, T., Lintu, N., Väistö, J., Leppänen, P. H., Laaksonen, D.E., Lindi, V. & Lakka, T. A. 2014a. Associations of physical activity and sedentary behavior with academic skills. A follow-up study among primary school children. PlosS one 9(9), e107031. Haapala, E. A., Poikkeus, A. M., Tompuri, T., Kukkonen-Harjula, K., Leppänen, P. H., Lindi, V. & Lakka, T. A. 2014b. Associations of motor and cardiovascular performance with academic skills in children. Medicine and Science in Sports and Exercise 46 (5), 1016–1024. Halperin, J. M., Sharma, V., Greenblatt, E. & Schwartz, S. T. 1991. Assessment of the continuous performance test: Reliability and validity in a nonreferred sample. Psychological Assessment: A Journal of Consulting and Clinical Psychology 3 (4), 603–608. Hamilton, L. S., Stecher, B. M. & Yuan, K. 2012. Standards-based accountability in the United States: Lessons learned and future directions. Education Inquiry 3(2), 149–170. Hancox, R. J., Milne, B. J. & Poulton, R. 2005. Association of television viewing during childhood with poor educational achievement. Archives of Pediatrics & Adolescent Medicine 159 (7), 614–618. Hanson, S. L. & Kraus, R. S. 1998. Women, sports, and science: Do female athletes have an advantage? Sociology of Education 71, 93–110. Heaton, R. K., Chelune, G., Talley, J. L., Kay, G. & Curtiss, G. 1993. Wisconsin card sorting test manual. Psychological Assessment Resources Odessa, FL. Henry, L. A. & Bettenay, C. 2010. The assessment of executive functioning in children. Child and Adolescent Mental Health 15 (2), 110–119. Herzog, W. & Boomsma, A. 2009. Small-sample robust estimators of noncentralitybased and incremental model fit. Structural Equation Modeling 16 (1), 1–27. Hill, L., Williams, J. H., Aucott, L., Milne, J., Thomson, J., Greig, J., Munro, V. & MonWilliams, M. 2010. Exercising attention within the classroom. Developmental Medicine & Child Neurology 52 (10), 929–934. Hillman, C. H., Buck, S. M., Themanson, J. R., Pontifex, M. B. & Castelli, D. M. 2009. Aerobic fitness and cognitive development: Event-related brain potential and task performance indices of executive control in preadolescent children. Developmental Psychology 45 (1), 114–129. Hillman, C. H., Castelli, D. M. & Buck, S. M. 2005. Aerobic fitness and neurocognitive function in healthy preadolescent children. Medicine and Science in Sports and Exercise 37 (11), 1967–1974. Hillman, C. H., Erickson, K. I. & Kramer, A. F. 2008. Be smart, exercise your heart: Exercise effects on brain and cognition. Nature Reviews Neuroscience 9 (1), 58– 65. Hillman, C. H., Kamijo, K. & Scudder, M. 2011. A review of chronic and acute physical activity participation on neuroelectric measures of brain health and cognition during childhood. Preventive Medicine 52, S21–S28. Hillman, C. H., Motl, R. W., Pontifex, M. B., Posthuma, D., Stubbe, J. H., Boomsma, D. I. & De Geus, E. J. 2006. Physical activity and cognitive function in a cross-section of younger and older community-dwelling individuals. Health Psychology 25 (6), 678–687. 76 Hillman, C. H., Pontifex, M. B., Raine, L. B., Castelli, D. M., Hall, E. E. & Kramer, A. F. 2009. The effect of acute treadmill walking on cognitive control and academic achievement in preadolescent children. Neuroscience 159 (3), 1044–1054. Hirvensalo, M. & Lintunen, T. 2011. Life-course perspective for physical activity and sports participation. European Review of Aging and Physical Activity 8 (1), 13– 22. Hogan, M., Kiefer, M., Kubesch, S., Collins, P., Kilmartin, L. & Brosnan, M. 2013. The interactive effects of physical fitness and acute aerobic exercise on electrophysiological coherence and cognitive performance in adolescents. Experimental Brain Research 229 (1), 85–96. Hopkins, M., Davis, F., Vantieghem, M., Whalen, P. & Bucci, D. 2012. Differential effects of acute and regular physical exercise on cognition and affect. Neuroscience 215, 59–68. Hu, L. & Bentler, P. M. 1999. Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal 6 (1), 1–55. Humes, G. E., Welsh, M. C., Retzlaff, P. & Cookson, N. 1997. Towers of Hanoi and London: Reliability of two executive function tasks. Assessment 4 (3), 249–257. Huotari, P. R., Nupponen, H., Laakso, L. & Kujala, U. M. 2010. Secular trends in aerobic fitness performance in 13–18-year-old adolescents from 1976 to 2001. British Journal of Sports Medicine 44 (13), 968–972. Iverson, J. M. 2010. Developing language in a developing body: the relationship between motor development and language development. Journal of Child Language 37 (2), 229–261. Jackson, L. A., Von Eye, A., Biocca, F. A., Barbatsis, G., Zhao, Y. & Fitzgerald, H. E. 2006. Does home internet use influence the academic performance of low-income children? Developmental Psychology 42 (3), 429–435. Jackson, L. A., Von Eye, A., Witt, E. A., Zhao, Y. & Fitzgerald, H. E. 2011. A longitudinal study of the effects of Internet use and videogame playing on academic performance and the roles of gender, race and income in these relationships. Computers in Human Behavior 27 (1), 228–239. Jeneson, A. & Squire, L. R. 2011. Working memory, long-term memory, and medial temporal lobe function. Learning & Memory 19 (1), 15–25. Johnson, J. G., Cohen, P., Kasen, S. & Brook, J. S. 2007. Extensive television viewing and the development of attention and learning difficulties during adolescence. Archives of Pediatrics & Adolescent Medicine 161 (5), 480–486. Kamijo, K., Pontifex, M., O'Leary, K., Scudder, M., Wu, C., Castelli, D. & Hillman, C. 2011. The effects of an afterschool physical activity program on working memory in preadolescent children. Developmental Science, 1-13. Kantomaa, M. T., Stamatakis, E., Kankaanpää, A., Kaakinen, M., Rodriguez, A., Taanila, A., Ahonen, T., Järvelin, M. & Tammelin, T. 2013. Physical activity and obesity mediate the association between childhood motor function and adolescents’ academic achievement. Proceedings of the National Academy of Sciences 110 (5), 1917–1922. Kantomaa, M. T., Tammelin, T. H., Demakakos, P., Ebeling, H. E. & Taanila, A. M. 2010. Physical activity, emotional and behavioural problems, maternal education and self-reported educational performance of adolescents. Health Education Research 25 (2), 368–379. 77 Karbach, J., Gottschling, J., Spengler, M., Hegewald, K. & Spinath, F. M. 2013. Parental involvement and general cognitive ability as predictors of domain-specific academic achievement in early adolescence. Learning and Instruction 23, 43–51. Kautiainen, S., Rimpela, A., Vikat, A. & Virtanen, S. M. 2002. Secular trends in overweight and obesity among Finnish adolescents in 1977-1999. International Journal of Obesity and Related Metabolic Disorders: Journal of the International Association for the Study of Obesity 26 (4), 544–552. Kohl III, H. W. & Cook, H. D. (Eds.) 2013. Educating the student body: Taking physical activity and physical education to school. National Academies Press. Koretz, D. M. & Hamilton, L. S. 2006. Testing for accountability in K-12. In Brennan R. L. (Ed.) Educational Measurement, 4th ed. Westport, CT: American Council on Education/Praeger, 531–578. Kim, H. P., Frongillo, E. A., Han, S., Oh, S., Kim, W., Jang, Y., Won, H., Lee, H. & Kim, S. 2003. Academic performance of Korean children in associated with dietary behaviours and physical status. Asia Pacific Journal of Clinical Nutrition 12 (2), 186–192. Kim, D. H. & So, W. Y. 2012. The relationship between daily Internet use time and school performance in Korean adolescents. Central European Journal of Medicine 7 (4), 444–449. Kirkorian, H. L., Wartella, E. A. & Anderson, D. R. 2008. Media and young children's learning. The Future of Children 18 (1), 39–61. Kramer, A. F. & Erickson, K. I. 2007. Capitalizing on cortical plasticity: Influence of physical activity on cognition and brain function. Trends in Cognitive Sciences 11 (8), 342–348. Kristjansson, A., Sigfusdottir, I. & Allegrante, J. 2010. Health behavior and academic achievement among adolescents: The relative contribution of dietary habits, physical activity, body mass index, and self-esteem. Health Education & Behavior 37 (1), 51–64. Kristjansson, A., Sigfusdottir, I., Allegrante, J. & Helgason, A. 2009. Adolescent health behavior, contentment in school, and academic achievement. American Journal of Health Behavior 33 (1), 69–79. Kühn, S., Gleich, T., Lorenz, R., Lindenberger, U. & Gallinat, J. 2013. Playing Super Mario induces structural brain plasticity: Gray matter changes resulting from training with a commercial video game. Molecular Psychiatry, 1–7. Kwak, L., Kremers, S., Bergman, P., Ruiz, J. & Rizzo, N. 2009. Associations between physical activity, fitness, and academic achievement. Journal of Pediatrics 155, 914–918. Lakshman, R., Elks, C. E. & Ong, K. K. 2012. Childhood obesity. Circulation 126 (14), 1770–1779. Landhuis, C. E., Poulton, R., Welch, D. & Hancox, R. J. 2007. Does childhood television viewing lead to attention problems in adolescence? Results from a prospective longitudinal study. Pediatrics 120 (3), 532–537. Lautenschlager, N. T., Cox, K. L., Flicker, L., Foster, J. K., van Bockxmeer, F. M., Xiao, J., Greenop, K. R. & Almeida, O. P. 2008. Effect of physical activity on cognitive function in older adults at risk for Alzheimer disease: A randomized trial. JAMA: The Journal of the American Medical Association 300 (9), 1027–1037. 78 LeBlanc, M. M., Martin, C. K., Han, H., Newton Jr, R., Sothern, M., Webber, L. S., Davis, A. B. & Williamson, D. A. 2012. Adiposity and physical activity are not related to academic achievement in school-aged children. Journal of Developmental & Behavioral Pediatrics 33 (6), 486–494. Levaux, M., Potvin, S., Sepehry, A. A., Sablier, J., Mendrek, A. & Stip, E. 2007. Computerized assessment of cognition in schizophrenia: Promises and pitfalls of CANTAB. European Psychiatry 22 (2), 104–115. Levine, L. E. & Waite, B. M. 2000. Television viewing and attentional abilities in fourth and fifth grade children. Journal of Applied Developmental Psychology 21 (6), 667–679. Lindner, K. J. 1999. Sport participation and perceived academic performance of school children and youth. Pediatric Exercise Science 11, 129–143. Lindner, K. J. 2002. The physical activity participation-academic performance relationship revisited: Perceived and actual performance and the effect of banding (academic tracking). Pediatric Exercise Science 14 (2), 155–169. Linebarger, D. L., Kosanic, A. Z., Greenwood, C. R. & Doku, N. S. 2004. Effects of viewing the television program Between the Lions on the emergent literacy skills of young children. Journal of Educational Psychology 96 (2), 297–308. Little, T. D., Cunningham, W. A., Shahar, G. & Widaman, K. F. 2002. To parcel or not to parcel: Exploring the question, weighing the merits. Structural Equation Modeling 9 (2), 151–173. Liu, Y., Wang, M., Tynjälä, J., Lv, Y., Villberg, J., Zhang, Z. & Kannas, L. 2010. Test-retest reliability of selected items of Health Behaviour in School-aged Children (HBSC) survey questionnaire in Beijing, China. BMC Medical Research Methodology 10 (1), 73. London, R. A. & Castrechini, S. 2011. A longitudinal examination of the link between youth physical fitness and academic achievement. Journal of School Health 81 (7), 400–408. Low, S. R. M. 1990. Influence of physical activity on concentration among junior highschool students. Perceptual and Motor Skills 70 (1), 67–74. Lowe, C. & Rabbitt, P. 1998. Test/re-test reliability of the CANTAB and ISPOCD neuropsychological batteries: Theoretical and practical issues. Neuropsychologia 36 (9), 915–923. Lu, L., Weber, H. S., Spinath, F. M. & Shi, J. 2011. Predicting school achievement from cognitive and non-cognitive variables in a Chinese sample of elementary school children. Intelligence 39 (2), 130–140. Luciana, M. 2003. Practitioner review: Computerized assessment of neuropsychological function in children: Clinical and research applications of the Cambridge Neuropsychological Testing Automated Battery (CANTAB). Journal of Child Psychology and Psychiatry 44 (5), 649–663. Luciana, M., Lindeke, L., Georgieff, M., Mills, M. & Nelson, C. A. 1999. Neurobehavioral evidence for working-memory deficits in school-aged children with histories of prematurity. Developmental Medicine & Child Neurology 41 (8), 521–533. Luciana, M. & Nelson, C. A. 2002. Assessment of neuropsychological function through use of the Cambridge Neuropsychological Testing Automated Battery: performance in 4- to 12-year-old children. Developmental Neuropsychology 22 (3), 595–624. 79 Luciana, M. & Nelson, C. A. 1998. The functional emergence of prefrontally-guided working memory systems in four-to eight-year-old children. Neuropsychologia 36 (3), 273–293. Maccoby, E. E. 1951. Television: Its impact on school children. Public Opinion Quarterly 15 (3), 421–444. Maeda, J. K. & Randall, L. M. 2003. Can academic success come from five minutes of physical activity? Brock Education 13 (1), 14–22. Malina, R. M. 2001. Physical activity and fitness: Pathways from childhood to adulthood. American Journal of Human Biology 13 (2), 162–172. Martínez-Gómez, D., Veiga, O. L., Gómez-Martínez, S., Zapatera, B., Martínez-Hernández, D., Calle, M. & Marcos, A. 2012. Gender-specific influence of health behaviors on academic performance in Spanish adolescents; The AFINOS study. Nutrición Hospitalaria 27 (3), 724–730. McClelland, M. M. & Cameron, C. E. 2011. Self-regulation and academic achievement in elementary school children. New Directions for Child and Adolescent Development 2011 (133), 29–44. McNaughten, D. & Gabbard, C. 1993. Physical exertion and immediate mental performance of sixth-grade children. Perceptual and motor skills 77 (3f), 1155– 1159. Miller, K. E., Melnick, M. J., Barnes, G. M., Farrell, M. P. & Sabo, D. 2005. Untangling the links among athletic involvement, gender, race, and adolescent academic outcomes. Sociology of Sport Journal 22 (2), 178–193. Milner, B. 1971. Interhemispheric differences in the localization of psychological processes in man. British Medical Bulletin 27 (3), 272–277. Mizuno, K., Tanaka, M., Fukuda, S., Imai-Matsumura, K. & Watanabe, Y. 2012. Divided attention of adolescents related to lifestyles and academic and family conditions. Brain and Development 35 (5), 435–440. Monti, J. M., Hillman, C. H. & Cohen, N. J. 2012. Aerobic fitness enhances relational memory in preadolescent children: The FITKids randomized control trial. Hippocampus 22 (9), 1876–1882. Moore, R. D., Drollette, E. S., Scudder, M. R., Bharij, A. & Hillman, C. H. 2014. The influence of cardiorespiratory fitness on strategic, behavioral, and electrophysiological indices of arithmetic cognition in preadolescent children. Frontiers in Human Neuroscience 8, 258. Mößle, T., Kleimann, M., Rehbein, F. & Pfeiffer, C. 2010. Media use and school achievement – boys at risk? British Journal of Developmental Psychology 28 (3), 69–725. Mountjoy, M., Andersen, L. B., Armstrong, N., Biddle, S., Boreham, C., Bedenbeck, H. P. B., Ekelund, U., Engebretsen, L., Hardman, K. & Hills, A. 2011. International Olympic Committee consensus statement on the health and fitness of young people through physical activity and sport. British Journal of Sports Medicine 45 (11), 839–848. Munasib, A. & Bhattacharya, S. 2010. Is the ‘Idiot's Box’ raising idiocy? Early and middle childhood television watching and child cognitive outcome. Economics of Education Review 29 (5), 873-883. Muthén, L. & Muthén, B. 1998–2012. Mplus User’s Guide. Seventh Edition. Los Angeles: Muthén & Muthén. 80 National core curriculum for basic education 2004. 2004. Helsinki: Finnish National Board of Education. Available: http://www.oph.fi/english/curricula_and_qualifications/basic_education. Assessed 6 May 2014. Neisser, U., Boodoo, G., Bouchard Jr, T. J., Boykin, A. W., Brody, N., Ceci, S. J., Halpern, D. F., Loehlin, J. C., Perloff, R. & Sternberg, R. J. 1996. Intelligence: Knowns and unknowns. American Psychologist 51 (2), 77–101. Nelson, M. C., Neumark-Stzainer, D., Hannan, P. J., Sirard, J. R. & Story, M. 2006. Longitudinal and secular trends in physical activity and sedentary behavior during adolescence. Pediatrics 118 (6), e1627–e1634. Nisbett, R. E., Aronson, J., Blair, C., Dickens, W., Flynn, J., Halpern, D. F. & Turkheimer, E. 2012. Intelligence: New findings and theoretical developments. American Psychologist 67 (2), 130–159. O'Dea, J. A. & Mugridge, A. C. 2012. Nutritional quality of breakfast and physical activity independently predict the literacy and numeracy scores of children after adjusting for socioeconomic status. Health Education Research 27 (6), 975–985. Ouakrim-Soivio, N. 2013. Toimivatko päättöarvioinnin kriteerit?: Oppilaiden saamat arvosanat ja Opetushallituksen oppimistulosten seuranta-arviointi koulujen välisten osaamiserojen mittareina. (In Finnish, abstract in English.) Raportit ja selvitykset 9. Opetushallitus. Owen, A. M., Downes, J. J., Sahakian, B. J., Polkey, C. E. & Robbins, T. W. 1990. Planning and spatial working memory following frontal lobe lesions in man. Neuropsychologia 28 (10), 1021–1034. Padilla-Moledo, C., Ruiz, J. R., Ortega, F. B., Mora, J. & Castro-Pinero, J. 2012. Associations of muscular fitness with psychological positive health, health complaints, and health risk behaviors in Spanish children and adolescents. Journal of Strength and Conditioning Research 26 (1), 167–173. Pagani, L. S., Fitzpatrick, C., Barnett, T. A. & Dubow, E. 2010. Prospective associations between early childhood television exposure and academic, psychosocial, and physical well-being by middle childhood. Archives of Pediatrics & Adolescent Medicine 164 (5), 425–431. Parsey, C. M. & Schmitter-Edgecombe, M. 2013. Applications of technology in neuropsychological assessment. The Clinical Neuropsychologist 27 (8), 1328–1361. Pate, R. R., Mitchell, J. A., Byun, W. & Dowda, M. 2011. Sedentary behaviour in youth. British Journal of Sports Medicine 45 (11), 906–913. Pate, R. R., O'Neill, J. R. & Lobelo, F. 2008. The evolving definition of "sedentary". Exercise and Sport Sciences Reviews 36 (4), 173–178. Pesce, C., Crova, C., Cereatti, L., Casella, R. & Bellucci, M. 2009. Physical activity and mental performance in preadolescents: Effects of acute exercise on free-recall memory. Mental Health and Physical Activity 2 (1), 16–22. Petersen, S. E. & Posner, M. I. 2012. The attention system of the human brain: 20 years after. Annual Review of Neuroscience 35, 73–89. Pontifex, M. B., Raine, L. B., Johnson, C. R., Chaddock, L., Voss, M. W., Cohen, N. J., Kramer, A. F. & Hillman, C. H. 2011. Cardiorespiratory fitness and the flexible modulation of cognitive control in preadolescent children. Journal of Cognitive Neuroscience 23 (6), 1332–1345. 81 Pontifex, M. B., Saliba, B. J., Raine, L. B., Picchietti, D. L. & Hillman, C. H. 2013. Exercise improves behavioral, neurocognitive, and scholastic performance in children with attention-deficit/hyperactivity disorder. The Journal of Pediatrics 162 (3), 543–551. Pontifex, M. B., Scudder, M. R., Drollette, E. S. & Hillman, C. H. 2012. Fit and vigilant: The relationship between poorer aerobic fitness and failures in sustained attention during preadolescence. Neuropsychology 26 (4), 407–713. Posner, M. I. & Rothbart, M. K. 2014. Attention to learning of school subjects. Trends in Neuroscience and Education 3 (1), 14–17. Puder, J., Marques-Vidal, P., Schindler, C., Zahner, L., Niederer, I., Bürgi, F., Ebenegger, V., Nydegger, A. & Kriemler, S. 2011. Effect of multidimensional lifestyle intervention on fitness and adiposity in predominantly migrant preschool children (Ballabeina): Cluster randomised controlled trial. BMJ: British Medical Journal 343. Raine, L. B., Lee, H. K., Saliba, B. J., Chaddock-Heyman, L., Hillman, C. H. & Kramer, A. F. 2013. The influence of childhood aerobic fitness on learning and memory. PloS One 8 (9), e72666. Rasmussen, M. & Laumann, K. 2013. The academic and psychological benefits of exercise in healthy children and adolescents. European Journal of Psychology of Education 28 (3), 945–962. Razel, M. 2001. The complex model of television viewing and educational achievement. The Journal of Educational Research 94 (6), 371–379. Reed, J., Einstein, G., Hahn, E., Hooker, S., Gross, V. & Kravitz, J. 2010. Examining the impact of integrating physical activity on fluid intelligence and academic performance in an elementary school setting: A preliminary investigation. Journal of Physical Activity & Health 7 (3), 343–351. Reed, J. A., Maslow, A. L., Long, S. & Hughey, M. 2013. Examining the impact of 45 minutes of daily physical education on cognitive ability, fitness performance, and body composition of African American youth. Journal of Physical Activity & Health 10 (2), 185–197. Reynolds, D. & Nicolson, R. I. 2007. Follow-up of an exercise-based treatment for children with reading difficulties. Dyslexia 13 (2), 78–96. Rhodes, S. M., Riby, D. M., Matthews, K. & Coghill, D. R. 2011. Attention-deficit/hyperactivity disorder and Williams syndrome: Shared behavioral and neuropsychological profiles. Journal of Clinical and Experimental Neuropsychology 33 (1), 147–156. Robbins, T. W., James, M., Owen, A. M., Sahakian, B. J., Lawrence, A. D., McInnes, L. & Rabbitt, P. M. 1998. A study of performance on tests from the CANTAB battery sensitive to frontal lobe dysfunction in a large sample of normal volunteers: Implications for theories of executive functioning and cognitive aging. Journal of the International Neuropsychological Society 4 (5), 474–490. Roberts, C. K., Freed, B. & McCarthy, W. J. 2010. Low aerobic fitness and obesity are associated with lower standardized test scores in children. The Journal of Pediatrics 156 (5), 711–718. Ruiz, J. R., Ortega, F. B., Castillo, R., Martín-Matillas, M., Kwak, L., Vicente-Rodríguez, G., Noriega, J., Tercedor, P., Sjöström, M. & Moreno, L. A. 2010. Physical activity, fitness, weight status, and cognitive performance in adolescents. The Journal of Pediatrics 157 (6), 917–922. 82 Sahakian, B. J., Owen, A. M., Morant, N. J., Eagger, S. A., Boddington, S., Crayton, L., Crockford, H. A., Crooks, M., Hill, K. & Levy, R. 1993. Further analysis of the cognitive effects of tetrahydroaminoacridine (THA) in Alzheimer's disease: Assessment of attentional and mnemonic function using CANTAB. Psychopharmacology 110 (4), 395–401. Sallis, J. F., McKenzie, T. L., Kolody, B., Lewis, M., Marshall, S. & Rosengard, P. 1999. Effects of health-related physical education on academic achievement: Project SPARK. Research Quarterly for Exercise and Sport 70 (2), 127–34. Salmon, J., Tremblay, M. S., Marshall, S. J. & Hume, C. 2011. Health risks, correlates, and interventions to reduce sedentary behavior in young people. American Journal of Preventive Medicine 41 (2), 197–206. Schmidt, M. E. & Vandewater, E. A. 2008. Media and attention, cognition, and school achievement. The Future of Children 18 (1), 63–85. Schmidt, M. E., Rich, M., Rifas-Shiman, S. L., Oken, E. & Taveras, E. M. 2009. Television viewing in infancy and child cognition at 3 years of age in a US cohort. Pediatrics 123 (3), e370–e375. Schumaker, J. F., Small, L. & Wood, J. 1986. Self-concept, academic achievement, and athletic participation. Perceptual and Motor Skills 62 (2), 387–390. Scudder, M. R., Federmeier, K. D., Raine, L. B., Direito, A., Boyd, J. K. & Hillman, C. H. 2014. The association between aerobic fitness and language processing in children: Implications for academic achievement. Brain and Cognition 87, 140–152. Shallice, T. & Shallice, T. 1982. Specific impairments of planning. Philosophical Transactions of the Royal Society of London. B, Biological Sciences 298 (1089), 199–209. Sharif, I. & Sargent, J. D. 2006. Association between television, movie, and video game exposure and school performance. Pediatrics 118 (4), e1061–e1070. Sharif, I., Wills, T. A. & Sargent, J. D. 2010. Effect of visual media use on school performance: A prospective study. Journal of Adolescent Health 46 (1), 52–61. Shephard, R. J. 1997. Curricular physical activity and academic performance. Pediatric Exercise Science 9, 113–126. Shephard, R. J. 2014. Physical activity of children and academic achievement. Medicine and Science in Sports and Exercise 46 (4), 840. Shi, X., Tubb, L., Fingers, S. T., Chen, S. & Caffrey, J. L. 2013. Associations of physical activity and dietary behaviors with children's health and academic problems. Journal of School Health 83 (1), 1–7. Shin, N. 2004. Exploring pathways from television viewing to academic achievement in school age children. The Journal of Genetic Psychology 165 (4), 367–382. Sibley, B. A. & Etnier, J. L. 2003. The relationship between physical activity and cognition in children: A meta-analysis. Pediatric Exercise Science 15 (3). Sigfusdottir, I. D., Kristjansson, A. L. & Allegrante, J. P. 2007. Health behaviour and academic achievement in Icelandic school children. Health Education Research 22 (1), 70–80. Silliker, S. A. & Quirk, J. T. 1997. The effect of extracurricular activity participation on the academic performance of male and female high school students. School Counselor 44 (4), 288–293. 83 Singh, A., Uijtdewilligen, L., Twisk, J. W., van Mechelen, W. & Chinapaw, M. J. 2012. Physical activity and performance at school: A systematic review of the literature including a methodological quality assessment. Archives of Pediatrics & Adolescent Medicine 166 (1), 49–55. Skotte, J., Korshøj, M., Kristiansen, J., Hanisch, C. & Holtermann, A. 2014. Detection of physical activity types using triaxial accelerometers. Journal of Physical Activity & Health 11 (1), 76–84. Smith, P. J., Need, A. C., Cirulli, E. T., Chiba-Falek, O. & Attix, D. K. 2013. A comparison of the Cambridge Automated Neuropsychological Test Battery (CANTAB) with “traditional” neuropsychological testing instruments. Journal of Clinical and Experimental Neuropsychology, 35 (3), 319–328. Smith, T. E. 1992. Time use and change in academic achievement: A longitudinal follow-up. Journal of Youth and Adolescence 21 (6), 725–747. So, W. 2012. Association between physical activity and academic performance in Korean adolescent students. BMC Public Health 12 (1), 258. Sollerhed, A. & Ejlertsson, G. 1999. Low physical capacity among adolescents in practical education. Scandinavian Journal of Medicine & Science in Sports 9 (5), 249– 256. Spence, I. & Feng, J. 2010. Video games and spatial cognition. Review of General Psychology 14 (2), 92–104. Spitzer, U. S. & Hollmann, W. 2013. Experimental observations of the effects of physical exercise on attention, academic and prosocial performance in school settings. Trends in Neuroscience and Education 2 (1), 1–6. Staiano, A. E., Abraham, A. A. & Calvert, S. L. 2012. Competitive versus cooperative exergame play for African American adolescents' executive function skills: Short-term effects in a long-term training intervention. Developmental Psychology 48 (2), 337–342. Staiano, A. E. & Calvert, S. L. 2011. Exergames for physical education courses: Physical, social, and cognitive benefits. Child Development Perspectives 5 (2), 93– 98. Stephens, L. J. & Schaben, L. A. 2002. The effect of interscholastic sports participation on academic achievement of middle level school students. NASSP: National Association of Secondary School Principals Bulletin 86 (630), 34–41. Stevens, T., To, Y., Stevenson, S. & Lochbaum, M. 2008. The importance of physical activity and physical education in the prediction of academic achievement. Journal of Sport Behavior 31 (4), 368–388. Strath, S. J., Kaminsky, L. A., Ainsworth, B. E., Ekelund, U., Freedson, P. S., Gary, R. A., Richardson, C. R., Smith, D. T. & Swartz, A. M. 2013. Guide to the assessment of physical activity: Clinical and research applications. A scientific statement from the American Heart Association. Circulation 128 (20), 2259–2279. Stroth, S., Kubesch, S., Dieterle, K., Ruchsow, M., Heim, R. & Kiefer, M. 2009. Physical fitness, but not acute exercise modulates event-related potential indices for executive control in healthy adolescents. Brain Research 1269, 114–124. Subrahmanyam, K., Greenfield, P., Kraut, R. & Gross, E. 2001. The impact of computer use on children's and adolescents' development. Journal of Applied Developmental Psychology 22 (1), 7–30. 84 Suoninen, A. 2013. Lasten mediabarometri 2012. 10–12-vuotiaiden tyttöjen ja poikien mediankäyttö. (In Finnish) Helsinki: Nuorisotutkimusverkosto/Nuorisotutkimusseura, verkkojulkaisuja 62. Swing, E. L., Gentile, D. A., Anderson, C. A. & Walsh, D. A. 2010. Television and video game exposure and the development of attention problems. Pediatrics 126 (2), 214–221. Tammelin, T., Laine, K., & Turpeinen, S. (Eds.) 2013. Oppilaiden fyysinen aktiivisuus [Physical Activity of School-aged children] (In Finnish, abstract in English). Research Reports on Sport and Health 272. Jyväskylä: LIKES – Foundation for Sport and Health Sciences. Tammelin, T. & Karvinen, J. (Eds.) 2008. Fyysisen aktiivisuuden suositus kouluikäisille 7–18-vuotiaille [Recommendatios for the physical activity of schoolaged children] (In Finnish, abstract in English). Helsinki: Opetusministeriö ja Nuori Suomi ry. Telama, R., Yang, X., Leskinen, E., Kankaanpää, A., Hirvensalo, M., Tammelin, T., Viikari, J. S. & Raitakari, O. T. 2014. Tracking of physical activity from early childhood through youth into adulthood. Medicine and Science in Sports and Exercise 46 (5), 955–962. Telford, R. D., Cunningham, R. B., Telford, R. M. & Abharatna, W. P. 2012. Schools with fitter children achieve better literacy and numeracy results: Evidence of a school cultural effect. Pediatric Exercise Science 24 (1), 45–57. Themane, M., Koppes, L., Kemper, H., Monyeki, K. & Twisk, J. 2006. The relationship between physical activity, fitness and educational achievement of rural South African children. Journal of Physical Education and Recreation 12 (1), 48–54. Tomkinson, G. R. & Olds, T. S. 2007. Secular changes in pediatric aerobic fitness test performance: The global picture. Medicine and Sport Science 50, 46–66. Tremarche, P. V., Robinson, E. M. & Graham, L. B. 2007. Physical education and its effect on elementary testing results. Physical Educator 64 (2), 58–64. Tremblay, M. S., Inman, J. W. & Willms, J. D. 2000. The relationship between physical activity, self-esteem, and academic achievement in 12-year-old children. Pediatric Exercise Science 12 (3), 312–323. Tremblay, M. S., LeBlanc, A. G., Janssen, I., Kho, M. E., Hicks, A., Murumets, K., Colley, R. C. & Duggan, M. 2011a. Canadian sedentary behaviour guidelines for children and youth. Applied Physiology, Nutrition, and Metabolism 36 (1), 59–64. Tremblay, M. S., LeBlanc, A. G., Kho, M. E., Saunders, T. J., Larouche, R., Colley, R. C., Goldfield, G. & Gorber, S. C. 2011b. Systematic review of sedentary behaviour and health indicators in school-aged children and youth. International Journal of Behavioral Nutrition and Physical Activity 8 (1), 98. Tremblay, M. S., Warburton, D. E., Janssen, I., Paterson, D. H., Latimer, A. E., Rhodes, R. E., Kho, M. E., Hicks, A., LeBlanc, A. G. & Zehr, L. 2011c. New Canadian physical activity guidelines. Applied Physiology, Nutrition, and Metabolism 36 (1), 36– 46. Troiano, R. P., Berrigan, D., Dodd, K. W., Mâsse, L. C., Tilert, T. & McDowell, M. 2008. Physical activity in the United States measured by accelerometer. Medicine and Science in Sports and Exercise 40 (1), 181–188. Trost, S. G. 2008. Physical education, physical activity, and academic performance in youth. Chronicle of Kinesiology and Physical Education in Higher Education 19 (3), 33–40. 85 Trudeau, F. & Shephard, R. J. 2010. Relationships of physical activity to brain health and the academic performance of schoolchildren. American Journal of Lifestyle Medicine 4 (2), 138–150. Trudeau, F. & Shephard, R. J. 2008. Physical education, school physical activity, school sports and academic performance. The International Journal of Behavioral Nutrition and Physical Activity 5, 10. Tuckman, B. W. & Hinkle, J. S. 1986. An experimental study of the physical and psychological effects of aerobic exercise on schoolchildren. Health Psychology 5 (3), 197–207. US Department of Health and Human Services & US Department of Health and Human Services 2008. Physical activity guidelines for Americans. Available at: http://www.health.gov/paguidelines/pdf/paguide.pdf. Assessed 29 April 2014. Van Dusen, D. P., Kelder, S. H., Kohl III, H. W., Ranjit, N. & Perry, C. L. 2011. Associations of physical fitness and academic performance among schoolchildren. Journal of School Health 81 (12), 733–740. Vaynman, S., Ying, Z. & Gomez-Pinilla, F. 2004. Hippocampal BDNF mediates the efficacy of exercise on synaptic plasticity and cognition. European Journal of Neuroscience 20 (10), 2580–2590. Verret, C., Guay, M. C., Berthiaume, C., Gardiner, P. & Beliveau, L. 2012. A physical activity program improves behavior and cognitive functions in children with ADHD: An exploratory study. Journal of Attention Disorders 16 (1), 71–80. Vindfeld, S., Schnohr, C. & Niclasen, B. 2009. Trends in physical activity in Greenlandic schoolchildren, 1994–2006. International Journal of Circumpolar Health 68 (1), 42–52. Voss, M. W., Chaddock, L., Kim, J. S., VanPatter, M., Pontifex, M. B., Raine, L. B., Cohen, N. J., Hillman, C. H. & Kramer, A. F. 2011. Aerobic fitness is associated with greater efficiency of the network underlying cognitive control in preadolescent children. Neuroscience 199 (29), 166–176. Wang, M. C., Haertel, G. D. & Walberg, H. J. 1994. What helps students learn? Educational Leadership 51 (4), 74–79. Weber, H. S., Lu, L., Shi, J. & Spinath, F. M. 2013. The roles of cognitive and motivational predictors in explaining school achievement in elementary school. Learning and Individual Differences 25, 85–92. Weis, R. & Cerankosky, B. C. 2010. Effects of video-game ownership on young boys’ academic and behavioral functioning: A randomized, controlled study. Psychological Science 21 (4), 463–470. Welk, G. J., Jackson, A. W., Morrow, J., James, R., Haskell, W. H., Meredith, M. D. & Cooper, K. H. 2010. The association of health-related fitness with indicators of academic performance in Texas schools. Research Quarterly for Exercise and Sport 81 (Supplement 2), 16S–23S. Wenger, E. 2000. Communities of practice and social learning systems. Organization 7 (2), 225–246. Wenger, E. 1998. Communities of practice: Learning, meaning, and identity. Cambridge University Press. Wigfield, A. & Cambria, J. 2010. Students’ achievement values, goal orientations, and interest: Definitions, development, and relations to achievement outcomes. Developmental Review 30 (1), 1–35. 86 Williams, P. A., Haertel, E. H., Haertel, G. D. & Walberg, H. J. 1982. The impact of leisure-time television on school learning: A research synthesis. American Educational Research Journal 19 (1), 19–50. Willoughby, T. 2008. A short-term longitudinal study of Internet and computer game use by adolescent boys and girls: Prevalence, frequency of use, and psychosocial predictors. Developmental Psychology 44 (1), 195–204. Wingfield, R. J., Graziano, P. A., McNamara, J. P. H. & Janicke, D. M. 2011. Is there a relationship between body mass index, fitness, and academic performance? Mixed results from students in a Southeastern United States elementary school. Current Issues in Education 14 (2). Winne, P. H. & Nesbit, J. C. 2010. The psychology of academic achievement. Annual Review of Psychology 61, 653–678. Wittberg, R., Cottrell, L. A., Davis, C. L. & Northrup, K. L. 2010. Aerobic fitness thresholds associated with fifth grade academic achievement. American Journal of Health Education 41 (5), 284–291. Wittberg, R. A., Northrup, K. L. & Cottrell, L. A. 2012. Children’s aerobic fitness and academic achievement: A longitudinal examination of students during their fifth and seventh grade years. American Journal of Public Health 102 (12), 2303– 2307. World Health Organization 2010. Global recommendations on physical activity for health. Available: http://whqlibdoc.who.int/publications/2010/97892415 99979_eng.pdf?ua=1. Assessed 29 April 2014. Wright, J. C., Huston, A. C., Murphy, K. C., St Peters, M., Piñon, M., Scantlin, R. & Kotler, J. 2001. The relations of early television viewing to school readiness and vocabulary of children from low-income families: The early window project. Child Development 72 (5), 1347–1366. Wu, C. T. & Hillman, C. H. 2013. Aerobic fitness and the attentional blink in preadolescent children. Neuropsychology 27 (6), 642–653. Wu, C. T., Pontifex, M. B., Raine, L. B., Chaddock, L., Voss, M. W., Kramer, A. F. & Hillman, C. H. 2011. Aerobic fitness and response variability in preadolescent children performing a cognitive control task. Neuropsychology 25 (3), 333–341. Yu, C., Chan, S., Cheng, F., Sung, R. & Hau, K. 2006. Are physical activity and academic performance compatible? Academic achievement, conduct, physical activity and self-esteem of Hong Kong Chinese primary school children. Educational Studies 32 (4), 331–341. Zimmerman, F. J. & Christakis, D. A. 2005. Children’s television viewing and cognitive outcomes: A longitudinal analysis of national data. Archives of Pediatrics & Adolescent Medicine 159 (7), 619–625. Zimmerman, F. J. & Christakis, D. A. 2007. Associations between content types of early media exposure and subsequent attentional problems. Pediatrics 120 (5), 986–992. Åberg, M. A. I., Pedersen, N. L., Torén, K., Svartengren, M., Bäckstrand, B., Johnsson, T., Cooper-Kuhn, C. M., Åberg, N. D., Nilsson, M. & Kuhn, H. G. 2009. Cardiovascular fitness is associated with cognition in young adulthood. Proceedings of the National Academy of Sciences 106 (49), 20906–20911. 87 88 APPENDICES 89 90 APPENDIX 1. Summary of studies examining the association of physical activity with academic performance and cognitive functions published in 2008 and after. Authors publication year Age and number of subjects, design, year, country (if mentioned) Measurement of PA and fitness, academic achievement and cognitive functions Main results Ardoy et al. 2014. 13 y, n=67, a 4month group-RCT: control group receiving 2 usual PE lessons, experimental group #1 receiving 4 usual PE lessons and experimental group #2 receiving 4 high intensity PE lessons, 2007, Spain. Academic achievement with grades in the core subjects (Math and Language), average score of others subjects, average score of all subjects and average score excluding PE, cognitive functions with the M (medium) version of the Spanish Overall and Factorial Intelligence Test (IGFM). Experimental group #2 improved significantly in average academic achievement (except language achievement), nonverbal and verbal abilities, abstract reasoning, spatial ability, numerical ability and verbal reasoning, compared to control group or experimental group #1. There were no differences between experimental group #1 and control group. Best et al. 2012. 6─10 y, n=33, a 2 x 2 within-subject experimental design: children’s cognitive functions were assessed after PA (physically active video games versus sedentary video activities) and cognitive engagement (challenging and interactive video games versus repetitive video activities), Georgia. Inhibitory control with a modified flanker task: the Child Attention Network Test (ANT-C). Children’s speed to resolve interference from conflicting visuospatial stimuli was enhanced after PA (exergaming), compared to sedentary activities. Cognitive engagement had no effect on task performance. Blom et al. 2011. Grades 3–8, n=2992, cross-sectional, 2007–2008, US. Physical fitness with FITNESSGRAM® (composite score of PACER, curl-ups, push-ups, trunk lifts, sit and reach test, and skinfold/BMI), academic achievement with the Mississippi Curriculum Test (MCT2) for language arts and mathematics. Positive correlation between overall physical fitness and language arts and math achievements. Booth et al. 2014. 11 y at baseline, n=4755, the Avon Longitudinal Study of Parents and Children: 2002– 2003→2004– 2005→2007–2010, UK. Objectively measured MVPA with Actigraph AM 7164 2.2 accelerometer (7 days, at least 3 days of 10 h valid data), academic achievement with nationally administered school assessments in mother tongue, math and science. MVPA predicted increased performance in mother tongue in both sexes at the ages of 11, 13 and 16. MVPA predicted increased performance in math in both sexes at the age of 16, but not at the age of 11 or 13. For females, the MVPA predicted increased science scores at 11 and 16, but not for males. 91 APPENDIX 1. Continued. Buck, Hillman & Castelli 2008. 7─12 y, n=74, cross-sectional, US. CRF with FITNESSGRAM® (PACER), inhibitory control with the Stroop task (word, colour, colour-word). Greater CRF was associated with better performance in each of the three conditions in the Stroop task. Budde et al. 2008. 13─16 y, n=115, children’s attention performance were measured after a regular school lesson (pre-test) and after 10 min. of coordinative exercise (experimental group) or of a normal sport lesson (control group), Germany. Attention and concentration with the d2-test. Both groups had enhanced attention and concentration performance with a significantly higher progression in the experimental group. Castelli et al. 2011. 8.8 y, n=59, crosssectional: children participated in a 9month physical activity programme during which their PA was measured, US. PA with Polar heart-rate monitors and time below, time at, and time above the target heart rate were recorded, CRF with maximal treadmill test using a modified Balke protocol, cognitive functions with the Stroop Color-Word Test and the Comprehensive Trail-Making Test. CRF has no correlation with cognitive functions. Mean time above the target hearth zone was associated with high executive function demand (Stroop Color-Word & Trail-Making part B). Carlson et al. 2008. From kindergarten to 5th grade, n=5316, the Early Childhood Longitudinal Study, fall 1998→ spring 2004, US. Classroom teachers measured the time spent in PE (minutes per week), academic achievement with standardized tests for math and reading. The higher amount of PE was positively associated with math and reading achievement in girls, but not in boys. Chaddock et al. 2010. 9─10 y, n=49, crosssectional, US. CRF with maximal treadmill test using a modified Balke protocol, memory with item and relational memory paradigm. CRF had a positive association with relational memory task performance, but no association with item memory task performance. Chaddock et al. 2010. 9─10 y, n=55, crosssectional, US. CRF with maximal treadmill test using a modified Balke protocol, inhibitory control with a modified version of the Eriksen flanker task. More fit children had less percent interference compared to less fit children. CRF fitness had no association with flanker accuracy or reaction time. Chaddock et al. 2011. 9─10 y, n=46, crosssectional, US. CRF with maximal treadmill test using a modified Balke protocol, memory performance with a modified version of a memory task developed by Henke et al. CRF fitness had a positive association with relational memory accuracy, but had no association with non-relational memory accuracy. 92 APPENDIX 1. Continued. Chaddock et al. 2012. 9─10 y, n=32, crosssectional, US. CRF with maximal treadmill test using a modified Balke protocol, inhibitory control with a modified version of the Eriksen flanker task. There were no group differences in congruent accuracy. During incongruent trials, only more fit children maintained accuracy across the blocks. There were no group differences in reaction times. Chaddock et al. 2012. 9─10 y, n=32, longitudinal: one year follow-up, US. CRF with maximal treadmill test using a modified Balke protocol, inhibitory control with a modified version of the Eriksen flanker task. More fit children had higher accuracy than less fit children across compatibility task conditions and test sessions. More fit children maintained accuracy across compatible and incompatible task conditions, whereas less fit children showed lower accuracy in the incompatible condition relative to the compatible condition. More fit children also had faster reaction times at followup. Chaddock et al. 2012. 8─10 y, n=26, crosssectional, US. CRF with maximal treadmill test using a modified Balke protocol, cognition with Virtual reality street crossing paradigm. More fit children maintained performance in street-crossing across all 3 conditions, whereas less fit children showed decreased performance when on the phone, relative to being undistracted or listening to music. ChaddockHeyman et al. 2013. 8–9 y, n=23, a RCT: 60+ minutes of PA 5 days per week for 9 months, US. Inhibitory control with a modified flanker task, including three task condition: neutral, incongruent, and NoGo trials. Children who participated in the PA programme improved their performance in the inhibitory control task at the level of young adults at post-test, while the control group still differed from the young adults. Chih et al. 2011. 11–12 y, n=476, cross-sectional, 2006–2010, China. Physical fitness measured with sit and reach test, bentknee sit-ups for one minute, 800m run/walk and standing long jump, academic achievement in math and mother tongue. Physical fitness variables were not associated with academic achievement. PE performance was positively associated with academic achievement. 93 APPENDIX 1. Continued. Chomitz et al. 2009. Grades 4,6,7,8 (mean age 11.7 y), n=1478, cross-sectional, 2004–2005, US. Physical fitness with five tests adapted from the Amateur Athletic Union (AAU) and FITNESSGRAM® (composite score of an endurance cardiovascular test, an abdominal strength test, a flexibility test, an upper-body strength test, and an agility test), the Massachusetts Comprehensive Assessment System (MCAS) for math and English. Overall physical fitness was positively associated with passing the MCAS English test and the MCAS Mathematics test. Crova et al. 2014. 9–10 y, n=70, a RCT: 6-month enhanced PE programme including cognitively demanding (open skill) activities or curricular physical education only, Italy. The random number generation (RNG) task measuring inhibitory control and working memory. Overweight children participating in cognitively demanding activities improved their inhibitory control but not working memory ability, compared to lean children and children who participated in curricular PE. More fit children had better inhibitory control but not working memory performance, compared to less fit children. Davis & Cooper 2011. 7–11 y, n=170, crosssectional, 2003– 2006, US. CRF fitness with a graded treadmill test (Modified Balke, Protocol for Poorly Fit Children), academic achievement with the WoodcockJohnson Tests of Achievement III: The Broad Reading and Broad Mathematics clusters, cognitive functions with the Cognitive Assessment System (CAS) consisting of 4 sub-categories: Planning, Attention, Simultaneous, and Successive. Peak VO2 and treadmill time were positively associated with math and reading test. Peak VO2 and treadmill time were positively associated with performance in Planning and Attention categories. Peak VO2 and treadmill time were not associated with performance in Simultaneous or Successive categories. Davis et al. 2011. 7–11 y, n=171, a RCT: exercise programme adding 20 or 40 minutes of aerobic exercise per day for 13 weeks, 2003– 2006, US. Academic achievement with the Woodcock-Johnson Tests of Achievement III: The Broad Reading and Broad Mathematics clusters, cognitive functions with the Cognitive Assessment System (CAS) consisting of 4 sub-categories: Planning, Attention, Simultaneous, and Successive and Antisaccade task. There was a dose-response benefit of PA on mathematics achievement, planning ability and inhibition. PA was not associated with reading achievement or attention, simultaneous, and successive performance. 94 APPENDIX 1. Continued. Donelly. et al. 2009. 7–9 y (at baseline), n=1527, a threeyear, cluster-RCT: 90 minutes of MVPA was combined with learning objectives in academic lessons during school week, 2003→2006, US. Academic achievement with the Wechsler Individual Achievement Test (2nd Edition) for reading, writing, mathematics and oral language skills. Academic achievement scores for composite, reading, spelling and math significantly improved from baseline to three years in children participating in intervention, compared to control group. Drollette et al. 2012. 9─11 y, n=36, a within-subjects counterbalanced design: children’s cognitive performance were measured before, during or after walking or seated rest, US. Inhibitory control and working memory with a modified flanker task and a modified spatial n-back task. After moderately intense aerobic walking, children’s performance in inhibitory control improved but working memory, compared to seated rest. There were no changes in task performance during actively walking or at seated rest in both tasks. Drollette et al. 2014. 8─10 y, n=40, children categorized by higher- and lowertask performance completed inhibitory control task before and after 20 min. of treadmill walking and seated rest, US. Inhibitory control with a modified flanker task. Following exercise, higher-performers maintained accuracy compared to seated rest, whereas lower-performers demonstrated improvements in inhibitory control accuracy following exercise. Duncan & Johnson 2014. 8–11 y, n=18, a repeated measures design: children completed academic achievement test following rest and two different intensity exercises, UK. 20 minutes of rest, 20 minutes on a cycling ergometer at 50% HRR, and 20 minutes on a cycling ergometer at 75% HRR, academic achievement with the Wide Range Achievement Test (WRAT 4). High and moderate intensity exercise improved spelling. Moderate intensity exercise improved reading. Exercise had no effect on sentence comprehension, but attenuated arithmetic. Edwards, Mauch & Winkelman 2011. 11–13 y, n=800, cross-sectional, 2005, US. PA with Youth Risk Behavior Surveillance Survey, physical fitness with FITNESSGRAM® (aerobic capacity with the mile run, muscle strength with push-ups and curl-ups), academic achievement with measures of Academic Progress (MAP) standardized tests for math and reading. Vigorous PA, but not moderated PA, CRF and participation in sports teams were positively associated with MAP math scores. Vigorous PA and moderate PA was positively associated with MAP reading scores. Increased CRF or participation in sports teams were not associated with MAP reading scores. Curl-ups and push-ups were not associated with MAP math or reading scores. 95 APPENDIX 1. Continued. Ericsson 2008. Grades 1-3, n=251, controlled intervention: intervention groups with 5 PE lessons and 1 motortraining lesson per week, control group with 2 PE lessons per week, 1999─2002, Sweden. Academic achievement with the national tests for mother language and mathematics, and reading development test, attention with Conners’ abbreviated questionnaire filled in by teacher. Children participating in intervention outperformed control children in reading development, mother language and mathematic tests. Intervention children had better teacher-reported attention (attention/ hyperactivity and impulse control, attention in total) in school year 2, but the difference were small and did not remain in school year 3. EvelandSayers et al. 2009. Grades 3─5 (mean age 9.7 y), 134, cross-sectional, 2005, US. CRF with one-mile run time, muscular fitness with combination of curl-ups and sitand-reach test, academic achievement with the TerraNova achievement test for math and reading/language arts. A negative association between one-mile run times and math scores in girls (also for total group). Negative association between one-mile run times and reading scores in girls. A positive relationship between muscular fitness and math scores only in total. No relationship between muscular fitness and reading/language arts. Fisher et al. 2011. 6.2 y, n=64, a RCT: an experiment group with 2 hours of aerobic exercise lessons per week for 10 weeks and a control group with 2 hours of standard physical education lessons, UK. Cognitive function with the Cognitive Assessment System (CAS) consisting of four subcategories: Planning, Attention, Simultaneous, and Successive, the Cambridge Neuropsychological Test Battery (CANTAB) with tests SSP and SWM, measuring working memory, the Attention Network Test (ANT) measuring inhibition. Children participating in intervention outperformed control children in working memory and inhibition measures (SSP, SWM, ANT accuracy) at posttest, but did not have better performance in other cognitive tests measuring planning, attention, simultaneous and successive. Fox et al. 2010. 14.9 y, n=4746, cross-sectional, 1998–1999, US. PA with self-reported PA and sport team participation, academic achievement with selfreported GPA. For high-school girls, PA and sports team participation were each independently associated with higher GPA. For highschool boys, only sports team participation was independently associated with higher GPA. For middle-school girls, PA (not sports team participation) were associated with higher GPA. For middleschool boys, PA and sports team participation were associated with higher GPA. 96 APPENDIX 1. Continued. Gallotta et al. 2012. 8─11 y, n=138, children’s attention abilities were tested before and immediately after school curricular lesson, traditional PE lesson, and coordinative PE lesson. Attention performance with the d2 Test. Each exertion type led to enhanced attention performance. Coordinative PE lessons led to a lower improvement in attentional performances, compared with both traditional physical education lessons and school curricular lessons. Haapala et al. 2014a. 6–8 y, n=186, prospective longitudinal: The Physical Activity and Nutrition in Children (PANIC) study and The First Steps Study, 2007─2009 and 2006─2011, Finland. PA with the PANIC Physical Activity Questionnaire, academic achievement with the nationally normed reading achievement battery (ALLU) for reading comprehension and fluency, and with a basic arithmetic test. Children who had more PA during recess and children who most often commuted actively to school had better reading fluency across grades 1–3. Children who engaged in organized sports had better arithmetic skills. Among boys, higher levels total PA and physically active school transportation were associated with reading fluency and reading comprehension. Among girls, total PA was positively associated with reading fluency and arithmetic skills only girls, whose parents had university level education, while the association was inverse in girls whose parents were less educated. Haapala et al. 2014b. 6─8 y, n=167, longitudinal: The Physical Activity and Nutrition in Children (PANIC) study and The First Steps Study, 2007─2009 and 2006─2011, Finland. CRF with a maximal cycle ergometer test, academic achievement with the nationally normed reading achievement battery (ALLU) for reading comprehension and fluency, and with a basic arithmetic test. CRF was not related to reading fluency, reading comprehension or arithmetic skills. Hill et al. 2010. 8─11 y, n=1224, a randomized crossover design: children in two groups received 30 minutes of classroom-based physical exercise for one week and no exercise during the other. Cognitive performance with paced serial addition, size ordering, listening span, digitspan backwards, digit-symbol encoding. Exercise benefit cognition: Group B receiving exercise during the second week gained benefits. However, group A receiving exercise during the firsts week did not gained benefits from exercise. 97 APPENDIX 1. Continued. Hillman et al. 2009. 9.4 y, n=38, crosssectional, US. CRF with FITNESSGRAM® (PACER), inhibitory control with the Eriksen flanker task. CRF has positive association with flanker task accuracy across conditions. Fitness has no association with reaction time. Hillman et al. 2009. 9.5 y, n=20, a withinsubjects design: children completed cognitive testing after rest and exercise, US. Inhibitory control with a modified Flanker task, academic achievement with the Wide Range Achievement Test - 3rd edition (WRAT3) for reading, spelling and arithmetic. After acute exercise, reading comprehension test performance increased. Acute exercise did not affect spelling or arithmetic. Acute treadmill walking did not affect response speed in flanker task. Response accuracy in incongruent flanker task trials increased after walking. Response accuracy in congruent condition was not affected. Hogan et al. 2013. 13─14 y, n=30, children classified as fit or unfit performed cognitive tasks after a 20 minutes of moderate intensity biking exercise and a period of relaxation, Germany. Physical fitness with a continuous-graded maximal exercise test, inhibitory control with Go/NoGo version of the Erikson flanker task. Acute PA or fitness level had no effect on reaction time in inhibitory control task. However, fit participants had significantly faster reaction times in the exercise condition in comparison with the rest condition and unfit participants showed significantly higher error rates for NoGo relative to Go trials in the rest condition, compared to exercise condition. Kamijo et al. 2011. 7–9 y, n=43, a RCT: 9-month afterschool physical activity programme including two hours per school day of aerobically demanding physical activities, US. Working memory with a modified Sternberg task. Response accuracy at post-test was greater than pre-test for the intervention group, but no such difference existed the for control group. There were no differences between groups in reaction times. Kantomaa et al. 2010. 15–16 y, n=7344, cross-sectional: The Northern Finland Birth Cohort, 2001– 2002, Finland. PA with self-reported MVPA, academic achievement with self-reported GPA. MVPA was positively associated with GPA. Kantomaa et al. 2013. 7–8 y → 15–16 y, n=8061, longitudinal: The Northern Finland Birth Cohort, 1992– 1004 → 2001–2002, Finland. Parent-reported motor function, self-reported PA, predicted cardiorespiratory fitness with a submaximal cycle ergometer test, teacherrated GPA. PA was associated with a higher GPA. Compromised motor function in childhood had a negative indirect effect on adolescents’ academic achievement via physical inactivity and obesity, not via CRF. 98 APPENDIX 1. Continued. Kristjánsson et al. 2009. 14–15 y, n=5810, cross-sectional: Youth in Iceland, 2000, Iceland. PA with self-reported PA, academic achievement with self-reported grades. PA was positively and moderately related to academic achievement. Kristjánsson, Sigfúsdóttir & Allegrante 2010. 14–15 y, n=6346, cross-sectional: Youth in Iceland, 2000, Iceland. PA with self-reported PA, academic achievement with self-reported grades. PA had direct positive association with academic achievement. Self-esteem was a weak mediator of the association of PA and academic achievement. Kwak et al. 2009. 16 y, n=232, crosssectional, Sweden. Objectively measured PA with an accelerometer (model WAM 7164) (4 days, at least 3 days of 10 h valid data), CRF with a maximal cycle ergometer test, academic achievement with scores from schools. Vigorous PA was positively associated with GPA only in girls. CRF was positively associated with GPA only in boys. LeBlanc et al. 2012. Grades 4–6 (mean age 10.4 y), n=261, cross-sectional, US. Objectively measured PA with an ActiGraph GT1M accelerometer (3 days, at least 2 days of 10h valid data), academic achievement with criterion-referenced tests for English/language arts, math, science and social studies. Objectively measured MVPA was not associated with English, math, science, or social studies. London & Castrechini 2011. Grades 4–6 at baseline, n=2735, longitudinal: the Youth Data Archive (YDA), 2002– 2003 → 2007–2008, US. Physical fitness with FITNESSGRAM® (composite score of aerobic capacity, body composition, abdominal strength and endurance, trunk extensor strength and endurance, upper body strength and endurance, and flexibility), academic achievement with the California standardized test (CST) in math and English/language arts. Overall physical fitness was positively associated with English and math in total. From 5th- to 7th-grade, overall physical fitness was positively associated with math, but not associated with English. For 7th-to 9th-grade, overall physical fitness was positively associated with math. For 7th-to 9th-grade, overall physical fitness was positively associated with English only in girls. MartínezGómez et al. 2012. 13–17 y, n=1825, cross-sectional, 2007–2008, Spain. PA with the Physician-based Assessment and Counselling for Exercise (PACE+) questionnaire, academic achievement with self-reported grades. In boys, there was no association between PA and academic performance. In girls, PA was associated with good math and language + math performance, but not language only. Monti, Hillman & Cohen 2012. 9.5 y, n=44, a RCT: a 9-month after-school aerobic exercise intervention including at least 70 min of MVPA every schoolday, US. Memory performance with a memory task inspired by Hannula et al. 2007 containing relational and item memory conditions. There were no differences in memory performance between children participating in the aerobic exercise programme and the control group. 99 APPENDIX 1. Continued. Moore et al. 2014. 9─10 y, n=40, crosssectional, US. CRF with maximal treadmill test using a modified Balke protocol, mathematics achievement with the Kaufman Test of Academic and Educational Achievement 2 (KTEA-2) and with an experimental arithmetic verification task, including small and large addition problems). There were no differences in mathematics achievement. More fit children had better performance in arithmetic verification tasks, including large addition problems, compared to less fit children. O’Dea & Mugridge 2012. Grades 3-7, n=824, cross-sectional, 2008, Australia. Self-reported PA, the standardized National Assessment Program for Literacy and Numeracy (NAPLAN). PA was not associated with literacy scores. PA was positively associated with math scores in boys, but not in girls. Padilla-Moledo et al. 2012. 6─17.9 y, n=690, cross-sectional, Spain. Muscular fitness with standing long jump test and basketball throwing test, self-reported academic performance compared with those of their classmates. Low muscular fitness was associated with low academic performance. Perce et al. 2009. 11─12 y, n=52, children’s cognitive functions were measured right after two physical education lessons (aerobic circuit training or team games) and baseline session, Italy. Free-recall memory performance with a test involving free-recall of items from a 20item word list. The team games (not the circuit training) improved immediate recall scores in both primacy and recency portions, compared to the baseline session. Both team game and circuit training improved delayed recall scores in the recency portion (not in primacy), compared to the baseline session. Pontifex et al. 2011. Mean age 10.1 y, n=48, cross-sectional, US. CRF with maximal treadmill test using a modified Balke protocol, inhibition with a modified version of the Eriksen flanker task. More fit children maintained accuracy across compatible and incompatible task conditions, whereas less fit children showed lower accuracy in the incompatible condition relative to the compatible condition. There were no group differences in reaction times. Pontifex et al. 2012. 9─11 y, n=62, crosssectional, US. CRF with maximal treadmill test using a modified Balke protocol, inhibition with a modified version of the Eriksen flanker task. Less fit children had poorer overall accuracy compared to more fit children, with a greater number of errors of omission, as well as longer and more frequent sequential errors of omission. 100 APPENDIX 1. Continued. Puder et al. 2011. Mean age 5.1, n=652, cluster-randomized controlled single-blinded trial: multidimensional lifestyle intervention including four 45 minute sessions of physical activity a week, 2008─2009, Switzerland. Cognitive functions with tests of attention and spatial working memory. The intervention had no effect on cognitive abilities. Raine et al. 2013. 9─10 y, n=48, crosssectional, US. CRF with maximal treadmill test using a modified Balke protocol, memory and learning ability with a map task in which children had to learn the names of specific regions on a map, under 2 learning conditions in which they only studied versus a condition in which they were tested during study. In terms of initial learning, there were no differences in task performance between more fit and less fit children. During the retention session, more fit children outperformed less fit children when the study-only learning strategy was used. Reed et al. 2010. 9.5 y, n=155, a RCT: 90 minutes of physical activity was integrated into the core curricula over four months, children were measured only after the intervention, 2008, US. Academic achievement with the Palmetto Achievement Challenge Tests (PACT) measuring English/language arts, mathematics, science, and social studies, cognitive functions with the Standard Progressive Matrices (SPM) Test of Fluid intelligence (total score). Experimental group children outperformed control children in social studies mandated academic achievement test. There were no significant differences in English/language arts, math and science achievement tests between the experimental and control groups. PA participation was positively associated with fluid intelligence. Reed et al. 2013. Grades 2–8 (mean age 10.2/11.2 y), n=470, controlled intervention: sevenmonth intervention providing 45 minutes of daily physical education, 2009–2010, US. Cognitive functions with the Standard Progressive Matrices (SPM) fluid intelligence test (5 sets: A-E) and the Perceptual Speed Test (sections 1–3). In elementary school, boys (not girls) in experimental group improved fluid intelligence performance on section D more than controls. In middle school, girls in experimental group improved fluid intelligence performance on sections B, C, D and E more than controls, whereas boys in experimental group improved their fluid intelligence performance on section E more than boys in control group. In elementary school, girls (not) in experimental group improved perceptual speed on sections 2 and 3 more than controls. 101 APPENDIX 1. Continued. Roberts, Freed & McCarthy 2010. Grades 5, 7 and 9, n=1989, cross-sectional, 2002–2003, US. CRF with a one-mile run/walk test, Academic achievement with California Achievement Tests version 6 (CAT6) and California Standards Tests (CST) for math and reading (CAT) or math and language (CST). Low CRF was associated with lower scores in math, reading and language. Ruiz et al. 2010. 13─18.5 y, n=1820, cross-sectional, 2002─2002, Spain. Self-reported participation in sports, CRF with 20-minute shuttle run test, upper-body muscular strength with handgrip strength test, lower-body muscular strength with standing long jump test, cognitive functions with SRA Test of Educational Ability (TEA). Participation in sports was positively associated with cognitive performance (verbal, numeric, and reasoning abilities and an overall score). CRF and muscular fitness were not associated with cognitive performance. Scudder et al. 2014. 9─10 y, n=46, crosssectional, US. CRF with maximal treadmill test according to a modified Balke protocol, academic achievement with the Standard Progressive Matrices (WRAT3) for reading, spelling, and arithmetic, sentence processing with Neuroscan STIM2 software. More fit children had greater reading and spelling scores compared to less fit children. There were no differences in arithmetic scores. In sentence processing tasks, more fit children exhibited overall shorter response times and performed more accurately across all sentence types. Shi et al. 2013. 7–14 y, n=3708, cross-sectional, 2008–2009, US. PA with self-reported PA, academic achievement with self-reported academic problems. PA was associated with less academic problems. So 2012. Grades 7–12, n=75066, cross-sectional: Korea Youth Risk Behavior Webbased Survey (KYRBWS-V), 2009, Korea. PA with self-reported frequency of VPA, frequency of MPA and frequency of strengthening exercises, selfreported academic performance. VPA was positively associated with academic performance in boys, but not in girls. MPA was positively associated with academic performance in both boys and girls. Strengthening exercises were not associated with academic performance. 102 APPENDIX 1. Continued. Spitzer & Hollmann 2013. 12–13 y, experiment #1 n=44, experiment #2 n=88, 2 experimental observations: 3 extra exercise lessons were added to school week for 4 months, Germany. Academic achievement with academic grades in mathematics, German (mother tongue), and English (foreign language), the d2 test of attention. In experiment #1, children participating in extra exercise slightly improved German grades, while the control group children worsened their grades. There were no changes in math or English scores. In experiment #2, children participating in extra exercise improved math grades, while the control group children had worse math grades. There were no changes in German or English scores. In experiment #2, children who participated in extra exercise had higher performance in attention tests compared to control children, but not in experiment #1. Stevens et al. 2008. From kindergarten to 5th grade, n~3200, the Early Childhood Longitudinal Study, 1998–1999→2002, US. PA with parent-reported physical activity of their children, academic achievement with standardized test scores in math and reading. PA was positively associated with mathematics and reading achievements in boys and girls. Stroth et al. 2009. 13─14 y, n=35, a controlled cross-over study design: children classified as fit and unfit performed cognitive task after 20 minutes of cycling exercise and 20 minutes of a rest period, Germany. CRF with a continuous graded maximal exercise test, inhibitory control with a Go/NoGo version of the Eriksen flanker task. Neither CRF nor exercise had any association with inhibitory control performance. Telford et al. 2012. Grades 2 and 4, n=757, cross-sectional, Australia. Objectively measured PA with pedometers, CRF with 20 minutes multistage run, academic achievement with government literacy and math test scores. PA and CRF were not associated with reading scores. CRF, but not PA, was positively associated with numeracy scores. PA and CRF were positively associated with writing scores. Van Dusen et al. 2011. Grades 3–11, 254743, cross-sectional, 2008–2009, US. Physical fitness with FITNESSGRAM® (CRF with the mile run or PACER test, muscular fitness with curl-ups, trunk lift and push-ups, and flexibility with shoulder stretch or the sit-and-reach test.), academic achievement with Texas Assessment of Knowledge and Skills (TAKS™) for reading and math. All physical fitness variables were positively associated with math and reading. 103 APPENDIX 1. Continued. Vindfeld, Schnohr & Niclasen 2009. 11–17 y, n=1366, cross-sectional: the Health Behaviour in School-Aged Children (HBSC) study in Greenland, 2006, Greenland. PA with self-reported PA, academic achievement with self-reported academic achievement. PA was positively associated with overall self-reported academic achievement. Voss et al. 2011. 9─10 y, n=36, crosssectional, US. CRF with maximal treadmill test using a modified Balke protocol, inhibitory control with a modified version of the Eriksen flanker task. More fit children were more accurate for the incongruent condition but not the congruent condition, compared to less fit children. There were no differences in reaction times. Welk et al.2010. Grades 3–12, 1770– 6864, cross-sectional, 2007, USA. CRF with FITNESSGRAM® (the mile run or the Pacer), academic achievement with Texas Assessment of Knowledge and Skills (TAKS™). CRF fitness was positively associated with academic performance (overall knowledge and skills assessment). Wingfield et al. 2011. Grades 4─5, 132, cross-sectional, 2008─2009, US. Physical fitness with the President’s Challenge Physical Activity and Fitness Awards Program (composite score of curl-ups, shuttle run, one mile run/walk, pull-ups, flexed-arm hanging), academic achievement with the Florida Comprehensive Assessment Test (FCAT) for reading and math. Overall physical fitness was not associated with reading and math. Wittberg et al. 2010. Grade 5, n=1740, cross-sectional, 2006–2008, US. CRF with FITNESSGRAM® (the mile run or the Pacer), academic achievement with West Virginia standardized academic test scores (WESTEST) for reading/ language arts, math, science and social studies. CRF was positively associated with academic performance. Wittberg, Northrup & Cottrell 2012. Grade 5 at baseline, n=1725, longitudinal study, 2005–2006 → 2007–2008, 2006– 2007 → 2008–2009, 2007–2008 → 2009– 2010, US. CRF with FITNESSGRAM® (the one-mile run or the PACER), academic achievement with West Virginia standardized academic test scores (WESTEST) for reading/ language arts, math, science and social studies. Students whose CRF stayed in the “healthy” fitness zone had significantly higher academic scores than did students whose CRF stayed in the “needs improvement” zone. Wu et al. 2011. 8─11 y, n=48, crosssectional, US. CRF with maximal treadmill test using a modified Balke protocol, inhibitory control with a modified version of the Eriksen flanker task. More fit children had better response accuracy across task conditions than lower fit children. There were no differences in reaction times. 104 APPENDIX 1. Continued. Wu & Hillman 2013. 9─10 y, n=39, crosssectional, US. CRF with maximal treadmill test using a modified Balke protocol, attention performance with a modified attentional blink paradigm with the same experimental settings and methodology as those used by Slagter et al. (2007). Higher-fit children had better attentional task performance than lower fit children. Åberg et al. 2009. 15→18, 1221727, longitudinal: Swedish men born 1950 through 1976 and who were enlisted for military service between 1968 and 1994, Sweden. CRF with a cycle ergometer test. Isometric muscle strength with knee extension, elbow flexion, and hand grip, intelligence with 4 cognitive tests (logical performance test, verbal test of synonyms and opposites, test of visuospatial/geometric perception, and technical/mechanical skills including mathematical/physics problems). CRF (not muscular strength) was positively associated with intelligence. Changes in CRF between age 15 and 18 years predicted cognitive performance at the age of 18. Abbreviations: CRF, cardiorespiratory fitness; GPA, grade point average; HRR, heart rate reserve; MPA, moderate physical activity; MVPA, moderate to vigorous physical activity; PA, physical activity; PACER, Progressive Aerobic Cardiovascular Endurance Run; PE, physical education; RCT, randomized controlled trial; VPA, vigorous physical activity. 105 APPENDIX 2. Summary of the studies examining the association of screen time with academic performance and cognitive functions. Authors and publication year Age and number of subjects, design, year, country (if mentioned) Measurements of screen time, academic achievement and cognitive functions Main results Bittman et al. 2011. Younger cohort: 0−1 y and older cohort 4−5 y at the baseline, n=5107 and n=4983, longitudinal: the Longitudinal Study of Australian Children (LSAC): children were assessed at four time points during two years, 2004−2007, Australia. TV viewing, computer use and game console ownership with parent-reported 24 h diary kept during one weekday and one weekend day, language ability with the Peabody Picture Vocabulary Test (3rd edition) (PPVT-III) and with the Language and Literacy Academic Rating Scale (ARS). TV viewing was not associated with language skills. Computer use was positively associated with development of language skills. The ownership of game consoles was negatively associated with literature achievements in the older cohort. Borzekowski & Robinson 2005. Grade 3, n=348, cross-sectional, 1999−2000, US. Media environment with self-reported questionnaire, academic achievement with the Stanford Achievement Test for math, reading, and language arts. Having TV in the children’s bedroom was negatively associated with academic achievement. Computer access and use were positively associated with academic achievement. Drowak et al. 2007. 13.5 y, n=11, children were exposed to computer game or video film between 6 pm – 7 pm, German. Visual and verbal memory with Swets Test Services 4 to 5 hours before bedtime on each experimental day. Excessive video game playing (but not television viewing) resulted decreased verbal memory performance (but not visuospatial memory performance), compared to basal condition. Dye & Bavelier 2009. 7−22 y, n=131, cross-sectional, US. Video game play with selfreported questionnaire, attention abilities (including inhibitory control) with the Attentional Network Test (ANT). Action-game playing was associated with enhanced attention skills. 106 APPENDIX 2. Continued. Ennemoser & Schneider 2008. 6.4 y and 8.6 y, n1=165 and n2=167, longitudinal: children’s TV viewing and reading performance were assessed at five time points during four years, 1998→2001, German. Screen time with parentreported diaries over 7 days, reading comprehension with a reading comprehension test developed by Näslund (1990) (grades 1−2), a subtest of the Knuspel Reading Test (grades 3−4), a subtest of the General School Achievement Test (grades 2−3), and subtests of the General German Language Test and the Zurich Reading Comprehension Test (grade 5), decoding speed with the Wϋrzburg Silent Reading Test. TV viewing had a negative association with tests of reading speed and comprehension. Children classified as heavy TV viewers, especially having high amounts of entertainment programme viewing, showed lower progress in reading performance compared to medium and light viewers in both age-groups. Espinoza 2009. High school juniors, cross-sectional, US. TV viewing with determining the number of appliances in the house and the time of operation of each appliance, academic achievement with physics performance. The number of television sets at home and the number of hours that the televisions are on were negatively associated with student performance in physics. Ferguson et al. 2013. 10−17, n=333/n=143, cross-sectional/prospective 1-year longitudinal, US. Videogame playing with self-reported questionnaire, academic achievement with Wide Range Achievement Test-IV (WRAT) for math, visuospatial cognition with the Kaufman Brief Intelligence Test-II. Violent video game exposure was not associated with math achievement or visuospatial cognition in short-term or in longterm. Gentile & Walsh 2002. 2−17 y, n=527, cross-sectional, 1998, US. Family media habits and academic achievement with parent-reported questionnaire. Amount of TV viewing and TV in children’s bedroom were negatively associated with school performance. 107 APPENDIX 2. Continued. Haapala et al. 2014a. 6–8 y, n=186, prospective longitudinal: The Physical Activity and Nutrition in Children (PANIC) study and The First Steps Study, 2007─2009 and 2006─2011, Finland. Sedentary behaviour with the PANIC Physical Activity Questionnaire, academic achievement with the nationally normed reading achievement battery (ALLU) for reading comprehension and fluency, and with a basic arithmetic test. Sedentary behaviour related to academic skills in 1st grade were positively associated with reading fluency in grades 1–3. Among boys (not among girls), higher levels of sedentary behaviour related to academic skills and higher computer use and video game play were associated with better reading fluency and arithmetic skills, respectively. TV viewing had no association with academic skills. Among girls, high levels of total sedentary behaviour (including sedentary behaviours related to screen time, music, academic skills, arts, crafts, games, and sitting and lying for a rest) and sedentary behaviour related to music and arts, crafts and games was negatively associated with arithmetic skills. Hancox, Milne & Poulton 2005. 5→26 y, n=1037, longitudinal: children were followed from the birth to age of 26, children were born 1972 and 1973, New Zealand. TV viewing with parent- or self-reported questionnaires at the age of 5, 7, 9, 11, 13 and 15, the highest level of educational attainment with four-point scale. Television viewing in childhood (ages 5–11 years) and adolescence (ages 13 and 15 years) had negative associations with later educational achievement (at the age of 26). Jackson et al. 2006. 10−18 y, n=140, longitudinal: children’s Internet use was continuously recorded and academic performance many time measured during 16 months, 2000−2002, US. Internet use with automatic recordings, academic achievement with GPAs and the Michigan Educational Assessment Program (MEAP) tests for reading and mathematics. Frequent Internet use was associated with higher reading (but not math) test scores and higher GPAs 6, 12 and 16 months later. Jackson et al. 2011. 12 y, n=482, longitudinal: children’s Internet use and video game playing were assessed at baseline and one year later, US. Internet use and video game play with self-reported questionnaire, academic achievement with self-reported GPAs and the Wide Range Achievement Test (WRAT-3) for reading and math, visuospatial skills with the Wide Range Assessment of Visual Motor Abilities Section 2, Matching. Video game playing was associated with lower GPAs. Internet use was associated with better reading skills (for children with initially low or average reading skills) and GPAs, but not math skills. Video game playing was positively associated with visuospatial skills. 108 APPENDIX 2. Continued. Johnson et al. 2007. 14→16→22 y, n=678, longitudinal: participants’ TV viewing and attention and learning difficulties were assessed at four time points, 1983→1985−1986 →1991−1993 TV viewing with the Disorganizing Poverty Interview, attention and learning difficulties with the Diagnostic Interview Schedule for Children at the age of 14 and 16, and age-appropriate version of the same test for 22 yearsolds. Frequent TV viewing at the age of 14 predicted poor academic grades, failure to complete high school, negative attitudes towards school and poor homework completion, as well as longterm academic failure. Children who watched television 3 or more hours per day at the age of 14, had higher prevalence of frequent attention difficulties at the age of 16, compared to children who watched less television. →2001−2004, US. Kim & So 2012. 13−18 y, n=75066, cross-sectional, 2009, Korea. Academic achievement and Internet use with selfreported questionnaire. Children who used Internet three hours a day or less had higher school performance than children who never used the Internet, while children who used the Internet over three hours a day had weaker performance. Landhuis et al. 2007. 3 y at baseline, 1037, longitudinal: children’s TV viewing was assessed at the age of 5, 7, 9, 11 and attention problems at the age of 13 and 15, children were born 1972 and 1973, New Zealand. TV viewing with parentand self-reported questionnaire, attention problems with self-, parentand teacher-rated questionnaires (the age-appropriate Diagnostic Interview Schedule for Children, the Quay and Peterson Revised Problem Behavior Checklist and the Rutter Child Scale (Scale B for teachers)). Both childhood and adolescent television viewing were independently associated with attention problems in adolescence. Mizuno et al. 2013. Grades 7−9, n=158, cross-sectional, Japan. TV viewing with self-reported questionnaire, cognitive function with the kana pick-out test. The high amounts of television viewing were associated with a weaker ability to divide attention. Munasib & Bhattacharya 2010. 5−10 y, cross-sectional: data from the National Longitudinal Survey of Youth (NLSY), 1990, 2002, US. TV viewing with motherreported questionnaire, academic achievement with the Peabody Individual Achievement Test (PIAT) for math and reading. There were no association between television viewing and academic achievement after adjusting for socioeconomic determinants, parents’ TV viewing behaviours and parents’ role in monitoring children’s viewing. 109 APPENDIX 2. Continued. Mößle et al. 2010. Study 1. Grade 4, n=5529, cross-sectional, 2005, German. Media use and academic achievement with self-reported questionnaires. Study 1: The time spent playing computer or video games and watching TV, DVDs or videos was negatively associated with school achievement, including grades in mother tongue, science and math. Study 2: In terms of crosssectional associations, media use – especially playing computer games – was negatively associated with marks in mother tongue, foreign language and science, but not math scores at all measurement occasions in 3rd, 4th and 5th grades. In longitudinal analysis, daily computer game playing was negatively associated with academic achievement, but TV usage had only few negative associations with academic achievements. Study 3: The decline in school grades was smaller in intervention groups, compared to control group. Study 2. Grade 3, n=1157, longitudinal: children’s media use and academic achievement were assessed in four times, 2005→2006→2007 →2008, German. Study 3. Grade 3 at baseline, N=1020, a RCT: school-based media intervention to affect children’s media habits, 2006–2008, Germany. Ruiz et al. 2010. 13─18.5 y, n=1820, cross-sectional, 2002─2002, Spain. Self-reported time spent in TV viewing and playing video games, cognitive functions with SRA Test of Educational Ability (TEA) (verbal, numeric, and reasoning abilities and an overall score). TV viewing or playing video games were not associated with cognitive functions. Sharif & Sargent 2006. Grades 5−8, n=4508, cross-sectional, US. TV viewing and video game playing and academic achievement with self-reported questionnaires. Weekday television viewing was negatively associated with school performance. Videogame play was not associated with school performance. Sharif, Wills & Sargent 2010. 10−14 y, n=6486, longitudinal: children’s media exposure and academic achievement were assessed at baseline, 8,16 and 24 months later, 2003→2005, US. Screen time and academic achievement with telephone interview. Screen exposure was negatively associated with change in school achievement. 110 APPENDIX 2. Continued. Shin 2004. 6−13 y, n=1203, cross-sectional, 1997, US. TV viewing with parentreported 24 h diary kept during one weekday and one weekend day, academic achievement with the Woodcock–Johnson Revised Tests of Achievement for reading and math. Children who watched more television had weaker math and reading achievements. Swing et al. 2010. Younger: 6−12 y, n=1323, older: 18−32, n=210, cross-sectional, US. TV and video game exposure with self- and parent reported questionnaires, attention problems with teacher-reported questionnaire in childhood and with a composite of three self-report measures: the Adult ADHD Self-Report Scale (ASRS), the Brief SelfControl Scale (BSCS), and the Barratt Impulsiveness Scale (BIS-11) in the late adolescence/early adulthood. High amounts of TV viewing and video game playing was associated with attention problems. Weis & Cerankosky 2010. 6−9 y, n=64, a RCT: experimental group receiving a video game system at baseline and the control group receiving a videogame system after four months intervention. Academic achievement with The Woodcock–Johnson—III: Tests of Achievement (WJ–III) for reading, mathematics and written language, school and home behaviour (including attention and learning difficulties) with The Parent Rating Scale (PRS) and Teacher Rating Scale (TRS). At the post-test, children in the experimental group, who had higher amounts of video game play, got lower reading and writing scores, but not math scores, compared to control children. The ownership of a video game system or video game playing did not affect attention problems. Willoughby 2008. Grades 9−10 at baseline, n=1591, longitudinal: children’s Internet use, computer use and academic performance were assessed at baseline and 21 months later, Canada. Internet and computer use with self-reported questionnaire, academic performance with standardized scores for ratings of typical school grades. There were no association between computer game play and academic performance. The moderate use of Internet was associated with more positive academic performance, compared to nonuse or high use. Abbreviations: ADHD, attention deficit hyperactivity disorder; GPA, grade point average; RCT, randomized controlled trial. 111 ORIGINAL PUBLICATIONS I PHYSICAL ACTIVITY, SEDENTARY BEHAVIOR, AND ACADEMIC PERFORMANCE IN FINNISH CHILDREN by Heidi J. Syväoja, Marko T. Kantomaa, Timo Ahonen, Harto Hakonen, Anna Kankaanpää & Tuija H. Tammelin. 2013. Medicine and Science in Sports and Exercise 45(11), 2098–2104. Reproduced with kind permission by Wolters Kluwer Health Lippincott & Wilkins. Physical Activity, Sedentary Behavior, and Academic Performance in Finnish Children ¨ OJA1,2, MARKO T. KANTOMAA1,3, TIMO AHONEN2, HARTO HAKONEN1, ANNA KANKAANPA ¨A ¨ 1, HEIDI J. SYVA 1 and TUIJA H. TAMMELIN 1 Research Center for Sport and Health Sciences, LIKES—Research Center for Sport and Health Sciences, Jyva¨skyla¨, FINLAND; 2University of Jyva¨skyla¨, Jyva¨skyla¨, FINLAND; and 3Department of Epidemiology and Biostatistics, MRC-HPA Centre for Environment and Health, Imperial College London, London, UNITED KINGDOM ABSTRACT EPIDEMIOLOGY ¨ OJA, H. J., M. T. KANTOMAA, T. AHONEN, H. HAKONEN, A. KANKAANPA ¨A ¨ , and T. H. TAMMELIN. Physical Activity, SYVA Sedentary Behavior, and Academic Performance in Finnish Children. Med. Sci. Sports Exerc., Vol. 45, No. 11, pp. 2098–2104, 2013. Purpose: This study aimed to determine the relationships between objectively measured and self-reported physical activity, sedentary behavior, and academic performance in Finnish children. Methods: Two hundred and seventy-seven children from five schools in the Jyva¨skyla¨ school district in Finland (58% of the 475 eligible students, mean age = 12.2 yr, 56% girls) participated in the study in the spring of 2011. Self-reported physical activity and screen time were evaluated with questions used in the WHO Health Behavior in School-Aged Children study. Children_s physical activity and sedentary time were measured objectively by using an ActiGraph GT1M/ GT3X accelerometer for seven consecutive days. A cutoff value of 2296 counts per minute was used for moderate-to-vigorous physical activity (MVPA) and 100 counts per minute for sedentary time. Grade point averages were provided by the education services of the city of Jyva¨skyla¨. ANOVA and linear regression analysis were used to analyze the relationships among physical activity, sedentary behavior, and academic performance. RESULTS: Objectively measured MVPA (P = 0.955) and sedentary time (P = 0.285) were not associated with grade point average. However, self-reported MVPA had an inverse U-shaped curvilinear association with grade point average (P = 0.001), and screen time had a linear negative association with grade point average (P = 0.002), after adjusting for sex, children_s learning difficulties, highest level of parental education, and amount of sleep. Conclusions: In this study, self-reported physical activity was directly, and screen time inversely, associated with academic achievement. Objectively measured physical activity and sedentary time were not associated with academic achievement. Objective and subjective measures may reflect different constructs and contexts of physical activity and sedentary behavior in association with academic outcomes. Key Words: ACADEMIC ACHIEVEMENT, MODERATE-TO-VIGOROUS PHYSICAL ACTIVITY, SCREEN TIME, SCHOOL AGE, LEARNING I and youth (8,3,13,17,30). However, diverging results have also been reported (20,26,31), indicating a somewhat weak and inconsistent association between young people_s physical activity and academic performance. Most of the previous studies used self-reported measures of physical activity and academic performance, with only a few reporting objectively measured physical activity in association with teacher-rated educational outcomes. In addition, previous studies were conducted in various countries with varying educational systems, which makes comparing the results difficult. According to previous studies, media use (22,27), especially time spent viewing TV (10,11,15,29), playing videogames (14), and using the Internet (18), has a negative association with academic achievement in childhood. However, Borzekowski and Robinson (1), and Jackson et al. (14) reported a positive relationship between Internet/computer use and academic performance, and Munasib and Bhattacharya (24) reported no association between television viewing and academic achievement. Objectively measured sedentary time is a current topic of interest in physiology but has not been extensively studied from the psychological point of view. To our knowledge, no one has studied the association between objectively measured sedentary time and academic performance. t has been proposed that only 30%–40% of youth are sufficiently active according to current public health recommendations (9). In contrast, sedentary behavior, especially screen-based sedentary behavior, has increased during the last few decades, and today children spend 4–8 hIdj1 sedentary (25). Physical inactivity has been shown to be associated with higher levels of obesity, metabolic, and cardiovascular risk factors, depression symptoms, and lower physical fitness in children, whereas adequate physical activity may benefit them (23). In addition to these health benefits, physical activity may have a beneficial effect on academic performance in children Address for correspondence: Heidi Syva¨oja, M.Sc., Viitaniementie 15a 40720, Jyva¨skyla¨, Finland; E-mail: heidi.syvaoja@likes.fi. Submitted for publication November 2012. Accepted for publication April 2013. 0195-9131/13/4511-2098/0 MEDICINE & SCIENCE IN SPORTS & EXERCISEÒ Copyright Ó 2013 by the American College of Sports Medicine DOI: 10.1249/MSS.0b013e318296d7b8 2098 Copyright © 2013 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited. The aim of this study was to examine the associations between objectively measured and self-reported physical activity, sedentary behavior, and teacher-rated academic achievement in children. We hypothesized that physical activity is directly, and sedentary behavior inversely, associated with academic achievement in childhood. STUDY POPULATION AND METHODS PHYSICAL ACTIVITY AND ACADEMIC PERFORMANCE Medicine & Science in Sports & Exercised Copyright © 2013 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited. 2099 EPIDEMIOLOGY Participants. During spring 2011, 475 fifth and sixth graders from five schools in the Jyva¨skyla¨ school district in Finland were invited to participate in the study. Fifty-eight percent (N = 277) of 475 eligible children participated in the study. The children were given an information pack containing a leaflet for themselves, a letter for their parents/ guardians, and a consent form. Participation in the study was voluntary, and all participants and their parents were informed about their right to drop out of the study any time without a specific reason. Only children with a fully completed consent form (certificate of consent signed by a parent/guardian and the child) on the day of the first measurements were included in the study. The study was performed according to the principles of the Declaration of Helsinki and the Finnish legislation and was approved by the Ethics Committee of the University of Jyva¨skyla¨. Academic achievement. Academic achievement scores (grades in individual school subjects and grade point averages [GPA]) were provided by the education services of the city of Jyva¨skyla¨. Individual grades were assessed in the following school subjects: native language (in most cases Finnish or Swedish), first foreign language (started in grade 3), mathematics, physics/chemistry, biology, history, geography, religion or ethics, visual arts, music, and physical education. The grades refer to numerical assessment on a scale of 4–10, where 4 denotes a failure (US grade: F) and 10 denotes excellent knowledge and skills (US grade: A). The GPA were calculated as means of the individual grades and were used as a measure of academic achievement in the analysis. A Finnish GPA of 5.0–5.9 equals 1.0 in US GPA, 6.0–6.9 equals 2.0, 7.0–8.9 equals 3.0, and 9.0–10.0 equals 4.0. Self-reported physical activity and screen time. Children filled in a questionnaire concerning demographics, habits, amount of sleep, and so on. Self-reported physical activity and screen time were evaluated with questions used in the WHO Health Behavior in School-Aged Children study (6). Self-reported moderate-to-vigorous physical activity (MVPA) was measured with the following question: ‘‘Over the past 7 days, on how many days were you physically active for a total of at least 60 minutes per day?’’ The response categories were as follows: 0, 1, 2, I, 7 d. Before this question, the following short description about what kind of physical activity should be taken into count was found: ‘‘In the next question, physical activity is defined as any activity that increases your heart rate and makes you get out of breath some of the time.’’ Examples of MVPA were running, walking quickly, rollerblading, biking, dancing, skateboarding, swim- ming, snowboarding, cross-country skiing, soccer, basketball, and Finnish baseball. Test–retest agreement for self-reported MVPA has been very good (ICC = 0.82) (21). Self-reported screen time was measured with the following question: ‘‘About how many hours a day do you usually a) watch television (including videos), b) play computer or video games, or c) use a computer (for purposes other than playing games, for example, e-mailing, chatting, or surfing the Internet or doing homework) in your free time?’’ There were response options for weekdays and weekends. The test–retest agreement for watching television (ICC = 0.72–0.74) and for playing computer or video games (ICC = 0.54–0.69) has been substantial and for using a computer (ICC = 0.33–0.50) fair to moderate (21). Daily screen time averages were calculated by adding these three questions including weekdays and weekends together. Objectively measured physical activity and sedentary time. Children_s physical activity was measured objectively by using the ActiGraph GT1M/GT3X accelerometer with one vertical axel. Children wore the accelerometer on the right hip with an elastic waistband during waking hours for seven consecutive days. Bathing, swimming, and other water activity periods were excluded. To collect data, the ActiLife accelerometer software (ActiLife version 5; http://support.theactigraph.com/dl/ActiLife-software) was used. Epoch length was 10 s, and nonwearing time was 30 min. For data reduction and analysis, a customized software was used. A cutoff value of 2296 counts per minute was used for MVPA (12) and 100 counts per minute for sedentary time. Children were included in the analysis if they had valid data for at least 500 minIdj1 on two weekdays and on one weekend day. Objectively measured sedentary time was standardized with daily monitoring time, which allowed the children, who had worn the accelerometers for different amounts of time per day, to be compared. Potential confounders. The parent or the child_s main caregiver filled in a questionnaire concerning family background. The mother_s and father_s education, family income, marital status, and children_s learning difficulties were investigated. The highest level of parental education, which was calculated from the mother_s and father_s education, was categorized as 1 = tertiary level education and 0 = basic or upper secondary education. The marital status of the main carer was categorized as 1 = married or cohabiting and 0 = divorced or single/widow. Children_s learning difficulties were evaluated with the following question: ‘‘Does your child have any diagnosed learning difficulties?’’ (categorization, 1 = yes and 0 = no or do not know). Statistical analyses. The Statistical Package for the Social Sciences was used for the statistical analyses (SPSS, 2010, IBM SPSS Statistics 19 Core System User_s Guide; SPSS Inc., Chicago, IL). Logarithmic transformations were used for variables with skewed distributions. Because the distribution of self-reported MVPA was negatively skewed, the distribution was reflected and logarithmically transformed and then reflected again to restore the original order of the TABLE 1. Sample characteristics according to sex and all participants. Boys Age (yr) GPA (4–10)b Objectively measured MVPA (minIdj1)c Objectively measured sedentary time (%Idj1)d Self-reported screen time (hIdj1) Girls All Mean SD n Mean SD n Mean SD n Pa 12.2 8.07 59.9 39.6 3.82 0.7 0.72 22.3 3.5 1.96 123 122 95 95 121 12.2 8.37 56.3 40.8 3.49 0.6 0.57 17.1 3.1 1.86 154 153 125 125 154 12.2 8.24 57.9 40.3 3.63 0.6 0.66 19.5 3.3 1.91 277 275 220 220 275 0.765 G0.001 0.623 0.006 0.095 a P values for sex differences (t-test). The GPA (scale 4–10, where 10 is the highest GPA one can earn) was calculated from the grades in the following individual school subjects: native language (in most cases Finnish or Swedish), first foreign language (started in grade 3), mathematics, physics/chemistry, biology, history, geography, religion or ethics, visual arts, music, and physical education. c MVPA measured with the ActiGraph accelerometer using a cutoff value 2296 counts per minute. d Sedentary time measured by the ActiGraph accelerometer using a cutoff value 100 counts per minute and expressed as percentage of daily monitoring time (%Idj1). EPIDEMIOLOGY b variable (jln((max + 1) j y)). The distributions of selfreported screen time and objectively measured MVPA were positively skewed. To measure sex differences, the independent samples t-test was used. The cross-sectional associations among physical activity, sedentary behavior, and academic achievement were examined with ANOVA and linear regression analysis. For the ANOVA, children were divided into tertile groups (33% each) according to the amount of objectively measured MVPA (1, tertile e47.0 min; 2, tertile 47.1– 65.0 min; 3, tertile Q65.1 min) and sedentary time (1, tertile e38.4%; 2, tertile 38.5%–41.4%; 3, tertile Q41.5%). In addition, children were classified into groups according to the selfreported MVPA (1 = 0–2 dIwkj1, 2 = 3–4 dIwkj1, 3 = 5– 6 dIwkj1, 4 = 7 dIwkj1) and screen time (1 = 0.00–1.99 hIdj1, 2 = 2.00–2.99 hIdj1, 3 = 3.00–3.99 hIdj1, 4 = 4.00–4.99 hIdj1, 5 = Q5.00 hIdj1). Before the multiple regression, the Pearson_s correlation coefficients for continuous variables were calculated to estimate associations between single variables and GPA. To investigate whether the associations between self-reported or objectively measured MVPA and GPA are quadratic, quadratic terms were calculated using an equation x2 = ((x j mean(x)) (x j mean(x)), where x is the logarithmically transformed self-reported MVPA or logarithmically transformed objectively measured MVPA. After that we used enter approach for the multiple regression. To calculate change in R2, the variables of interest were added to the second block one by one, and all other variables of the model (potential confounders and other variables of interest) were added to the first block. The change in R2 for all variables of interest was calculated and tested for significance. To study whether the assumptions of the regression analysis were fulfilled, we examined the distribution of model residuals. Sample characteristics were summarized descriptively, using mean and SD values for continuous data and frequencies and percentages for categorical data. The level for statistical significance was determined as P G 0.05. The star symbols are used to illustrate statistical significance in the figures and tables (***P G 0.001, **P G 0.01, *P G 0.05). Seventy-six percent of parents were married or cohabiting. Seven percent of children had a diagnosed learning difficulty. On the basis of the teacher ratings, girls had higher GPA compared with boys (t228 = 6.26, P G 0.001) (Table 1). Boys reported MVPA for at least 60 minIdj1 more often than girls (t239 = 8.10, P = 0.049) (Table 2). On the basis of objective measurements, children had, on average, 58 min of MVPA per day, with no statistically significant difference between boys and girls (t162 = 5.34, P = 0.623) (Table 1). However, girls spent more of their waking hours sedentary than boys (t218 = 2.71, P = 0.006). On average, children reported 3.6 h of screen time per day, with no statistically significant difference between boys and girls (t273 = 1.11, P = 0.095) (Table 1). According to the ANOVA, a high level of self-reported MVPA was associated with a high GPA (F3, 268 = 6.56, P G 0.001) (Fig. 1a). Children who were physically active at least 60 minIdj1 for 5–6 dIwkj1 had the highest GPA (8.41), whereas children who were physically active 0–2 dIwkj1 had the lowest GPA (7.83). Screen time was inversely associated with GPA (F4, 268 = 7.08, P G 0.001) (Fig. 1b). Children who had less than 2 hIdj1 of screen time had the highest GPA (8.5), whereas children who had more than 5 hIdj1 of screen time had the lowest GPA (8.0). Objectively measured MVPA (F2, 216 = 0.17, P = 0.843) and sedentary time (F2, 216 = 0.46, P = 0.635) were not associated with GPA (Figs. 1b and 1c). According to Pearson_s correlation coefficients, self-reported screen time was negatively associated with GPA (P G 0.001) (Table 3), whereas objectively measured MVPA (P = 0.955) or sedentary time (P = 0.285) had no significant association with GPA. The quadratic term for self-reported MVPA was associated with GPA (P G 0.001) (Table 3), whereas the quadratic term for objectively measured MVPA was not TABLE 2. Self-reported MVPA. Self-Reported MVPA (Active Days per Week)a 0–1 2 3 4 5 6 7 n Pb RESULTS The mean age of the children was 12.2 yr, and 56% of the children were girls (Table 1). In 79% of families, the highest level of parental education was tertiary level education. 2100 Official Journal of the American College of Sports Medicine Boys (%) Girls (%) All (%) 5.0 1.7 7.4 19.8 19.0 13.2 33.9 123 1.3 7.2 13.1 15.0 22.2 22.9 18.3 154 2.9 4.7 10.6 17.2 20.8 18.6 25.2 277 0.049 The percentages of children who were physically active for at least 60 minIdj1 during 0–1, 2, 3, 4, 5, 6, or 7 dIwkj1 according to self-reports. b P values for the sex differences (t-test). a http://www.acsm-msse.org Copyright © 2013 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited. FIGURE 1—GPA with self-reported MVPA (A), self-reported screen time (B), objectively measured MVPA (C), and objectively measured sedentary time (D). Results from ANOVA. (P = 0.123). According to multiple regression analysis, selfreported MVPA had an inverse U-shaped curvilinear association with GPA (Fig. 2a, Table 3), and screen time had a linear negative association with schools_ grade average, after adjusting for sex, learning difficulties, the highest level of parental education, and amount of sleep (Fig. 2b, Table 3). The adjusted R2 for the model was 0.305. The regression model residuals were normally distributed. Summary of results. In this study, self-reported physical activity was directly, and screen time inversely, associated with academic achievement in children. Objectively measured physical activity and sedentary time were not associated with academic achievement. TABLE 3. Regression analysis of GPA. Regression Model b Children_s learning difficulties The highest level of parental educationc Amount of sleep (hIdj1) Sex (female) Self-reported MVPAd Quadratic self-reported MVPAe Self-reported screen time (hIdj1)f R2 Adjusted R2 N Pearson_s Correlations B (SE) Beta (SE) $R 2a j0.368*** 0.247*** 0.211*** 0.223*** 0.003 j0.247*** j0.276*** j0.697*** (0.139) 0.307** (0.087) 0.114* (0.052) 0.176* (0.072) 0.065 (0.061) j0.337** (0.104) j0.198** (0.062) j0.297*** (0.139) 0.207** (0.087) 0.129* (0.052) 0.144* (0.072) 0.067 (0.061) j0.199** (0.104) j0.193** (0.062) 0.004 0.035** 0.033** 0.328 0.305 212 The level of statistical significance: ***P G 0.001, **P G 0.01, *P G 0.05. a The change in R2 , self-reported MVPA, quadratic self-reported MVPA, and screen time were added to the model one by one. b Parental report of child_s diagnosed learning difficulties categorized as 1 = yes and 0 = no or do not know. c The highest level of parental education categorized as 1 = tertiary level education and 0 = basic or upper secondary education. d The distribution of self-reported MVPA was reflected and transformed logarithmically and reflect again to restore the original order of the variable (jln((max + 1)y)). e To measure quadratic association between self-reported physical activity and GPA, quadratic term was formed with the following equation: quadratic self-reported MVPA = ((x j mean(x)) (x j mean(x)), where x is the logarithmically transformed self-reported MVPA (jln((max + 1) j y)). f Distribution of self-reported screen time was transformed logarithmically (ln(y)). B, unstandardized coefficient; Beta, standardized coefficient; $R2, change in R2 . PHYSICAL ACTIVITY AND ACADEMIC PERFORMANCE Medicine & Science in Sports & Exercised Copyright © 2013 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited. 2101 EPIDEMIOLOGY DISCUSSION Self-reported physical activity and academic achievement. Our finding of a positive association between self-reported physical activity and academic achievement is consistent with previous studies that reported MVPA is associated with high levels of academic performance (13,17,30). In addition, Donelly et al. (8) reported that adding 90 min of MVPA to children_s school week resulted in improvements in academic achievement during the 3-yr intervention time. However, in our study, the relationship between self-reported MVPA and academic achievement was curvilinear. It seems that five to six times per week may be the optimal amount of MVPA from the perspective of academic achievement. It may be that some of the most active children spend time in physical activities at the expense of time devoted to homework. A positive association between physical activity and academic achievement may be due to effects of physical EPIDEMIOLOGY FIGURE 2—GPA with self-reported MVPA (A), self-reported screen time (B), objectively measured MVPA (C), and objectively measured sedentary time (D). Results from the regression analysis. activity on children_s cognitive function. Davis et al. (7) and Chaddock et al. (4) suggested that regular physical activity enhances executive functions. Likewise, Castelli et al. (2) and Kamijo et al. (16) in their intervention studies observed that increased physical activity had a positive influence on children_s executive functions and working memory. Objectively measured physical activity and academic achievement. In this study, objectively measured MVPA was not associated with children_s academic achievement. Our results support those of LeBlanc et al. (20), who reported that objectively measured MVPA was not associated with academic performance in children. In contrast, Kwak et al. (19) found that objectively measured vigorous physical activity was associated with academic achievement in girls, but not in boys. However, Kwak et al. (19) studied adolescents (15–16 yr), whereas we and LeBlanc et al. (20) studied children age 10–12 yr. Furthermore, in the present study, physical activity was measured during seven consecutive days, whereas LeBlanc et al. (20) and Kwak et al. (19) measured physical activity for 3 and 4 d, respectively. Explanations for the inconsistencies between selfreported and objectively measured MVPA in association with academic achievement. The inconsistency between the subjective and the objective measures of physical activity in association with academic achievement may be due to the difficulty estimating one_s overall physical activity. According to Corder et al. (5), 40% of inactive children age 10 yr old overestimated their physical activity compared with the objective measure. However, nowadays, skill-specific types of physical activities based on body movements are common among children but may not be seen in activity counts. For example, skateboarding is a skill-specific physical activity, which requires balance and agility, but hardly accumulates 2102 Official Journal of the American College of Sports Medicine activity counts. Therefore, self-reported MVPA may better illustrate different types of activity, including skill-specific activity, whereas accelerometer-measured MVPA mainly illustrates cardiovascular activity. Objective and subjective measures may reflect different constructs and contexts of physical activity in association with academic outcomes. Sedentary behavior and academic achievement. According to this study, self-reported screen time was inversely associated with academic performance. This finding is in line with previous studies (22,27), supporting the hypothesis of time displacement (28). Time displacement theory suggests that the time spent in front of the screen may simply displace time spent in other activities such as doing homework, reading books, or sleeping, which may independently affect academic performance. There may be certain dispositions in media use, especially intense and exciting sensations, which increase the desire for these kinds of experiences and are incompatible with concentrated effort reading and writing (28). In addition, attention difficulties, frequent failure to do homework, and negative attitude toward school have been reported to mediate the association between television viewing at the age of 14 yr and academic failure at the age of 22 yr (15). In the present study, objectively measured sedentary time was not associated with academic achievement. This might be because objective measures of sedentary time do not specify the elements of sedentary behavior. It is reasonable to suggest that some of the sedentary activities performed (e.g., doing homework and reading) benefit learning and academic achievement. Strengths and limitations. To our knowledge, this is the first study to examine the associations of objectively measured and self-reported physical activity on teacher-rated academic achievement. Our study sample is representative http://www.acsm-msse.org Copyright © 2013 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited. regarding physical activity, showing that the level of physical activity of Finnish school-age children is comparable with that reported in international results (6). Because of the crosssectional design, conclusions regarding causality of the observed relationships cannot be drawn. In addition, the time spent doing homework, reading, or performing other activities that may benefit academic achievement was not investigated, limiting more precise examination of sedentary behavior. Moreover, the detailed content of screen-based sedentary behavior was not assessed, limiting the interpretation of the results regarding screen-based sedentary behavior. Inconsistency in physical activity results may be due to the accelerometer method itself. The accelerometer does not measure swimming, cycling, or similar activities. Furthermore, 1 wk or less of objective measurement may not be long enough to depict children_s usual physical activity. Future research. More research, especially more randomized controlled trials and longitudinal studies, is needed to clarify the relationship between physical activity and academic achievement. Besides, more information about the factors that may explain the association between physical activity and academic performance is needed. In future studies, longer accelerometer wearing times and, preferably, several measurement periods during the school year should be considered. Moreover, the inconsistency between the subjective and objective measurements related to educational outcomes observed in this study offers an interesting viewpoint for future studies to consider. We recommend that future studies use subjective and objective measurements to examine physical activity, sedentary behavior, and academic achievement. Furthermore, methods should assess not only the amount but also different types of physical activity and sedentary behavior more precisely. CONCLUSIONS In conclusion, self-reported, but not objectively measured, physical activity and sedentary behavior were associated with academic achievement in children. Self-reported MVPA had an inverse U-shaped curvilinear relationship with GPA, whereas self-reported screen time had a linear negative relationship with GPA. This study was funded by Finnish Ministry of Education and Culture. The authors are grateful to Professor Asko Tolvanen for his advice in statistical analysis. There are no relationships, conditions, or circumstances that present potential conflict of interest. The results of this study do not constitute endorsement by the American College of Sports Medicine. REFERENCES PHYSICAL ACTIVITY AND ACADEMIC PERFORMANCE 11. Espinoza F. Using project-based data in physics to examine television viewing in relation to student performance in science. J Sci Educ Technol. 2009;18(5):458–65. 12. Evenson KR, Catellier DJ, Gill K, et al. Calibration of two objective measures of physical activity for children. J Sports Sci. 2008;24(14):1557–65. 13. Fox CK, Barr-Anderson D, Neumark-Sztainer D, et al. Physical activity and sports team participation: associations with academic outcomes in middle school and high school students. J School Health. 2010;80(1):31–7. 14. Jackson LA, Von Eye A, Witt EA, et al. A longitudinal study of the effects of Internet use and videogame playing on academic performance and the roles of gender, race and income in these relationships. Comput Hum Behav. 2011;27(1):228–39. 15. Johnson JG, Cohen P, Kasen S, et al. Extensive television viewing and the development of attention and learning difficulties during adolescence. Arch Pediatr Adolesc Med. 2007;161(5):480. 16. Kamijo K, Pontifex MB, O’Leary KC, et al. The effects of an afterschool physical activity program on working memory in preadolescent children. Dev Sci. 2011;14(5):1046–58. 17. Kantomaa MT, Tammelin TH, Demakakos P, et al. Physical activity, emotional and behavioural problems, maternal education and self-reported educational performance of adolescents. Health Educ Res. 2010;25(2):368–79. 18. Kim DH, So WY. The relationship between daily Internet use time and school performance in Korean adolescents. Cent Eur J Med. 2012;7(4):444–9. 19. Kwak L, Kremers SPJ, Bergman P, et al. Associations between physical activity, fitness, and academic achievement. J Pediatr. 2009;155:914–8. 20. LeBlanc MM, Martin CK, Han H, et al. Adiposity and physical activity are not related to academic achievement in school-aged children. J Dev Behav Pediatr. 2012;33(6):486–94. 21. Liu Y, Wang M, Tynja¨la¨ J, et al. Test–retest reliability of selected items of Health Behaviour in School-aged Children (HBSC) survey questionnaire in Beijing, China. BMC Med Res Methodol. 2010;10(1):73. Medicine & Science in Sports & Exercised Copyright © 2013 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited. 2103 EPIDEMIOLOGY 1. Borzekowski DLG, Robinson TN. The remote, the mouse, and the no. 2 pencil: the household media environment and academic achievement among third grade students. Arch Pediatr Adolesc Med. 2005;159(7):607. 2. Castelli DM, Hillman CH, Hirsch J, et al. FIT kids: time in target heart zone and cognitive performance. Prev Med. 2011;52:S55–9. 3. Centers for Disease Control and Prevention. The Association Between School-Based Physical Activity, Including Physical Education, and Academic Performance. Atlanta (GA): U.S. Department of Health and Human Services; 2010. p. 84. 4. Chaddock L, Hillman CH, Buck SM, et al. Aerobic fitness and executive control of relational memory in preadolescent children. Med Sci Sports Exerc. 2011;43(2):344–9. 5. Corder K, van Sluijs EMF, McMinn AM, et al. Perception versus reality: awareness of physical activity levels of British children. Am J Prev Med. 2010;38(1):1–8. 6. Currie C, Zanotti C, Morgan A, Currie D, De Looze M, Roberts C, Samdal O, Smith ORF, Barnekow V. Social Determinants of Health and Well-being among Young People. Health Behaviour in School-aged Children (HBSC) Study. International Report from the 2009/2010 Survey:129–138, 237–238. 7. Davis CL, Tomporowski PD, McDowell JE, et al. Exercise improves executive function and achievement and alters brain activation in overweight children: a randomized, controlled trial. Health Psychol. 2011;30(1):91–8. 8. Donnelly JE, Greene JL, Gibson CA, et al. Physical Activity Across the Curriculum (PAAC): a randomized controlled trial to promote physical activity and diminish overweight and obesity in elementary school children. Prev Med. 2009;49(4):336–41. 9. Ekelund U, Tomkinson G, Armstrong N. What proportion of youth are physically active? Measurement issues, levels and recent time trends. Br J Sports Med. 2011;45(11):859–65. 10. Ennemoser M, Schneider W. Relations of television viewing and reading: findings from a 4-year longitudinal study. J Educ Psychol. 2007;99(2):349. 27. Sharif I, Sargent JD. Association between television, movie, and video game exposure and school performance. Pediatrics. 2006; 118(4):e1061–70. 28. Sharif I, Wills TA, Sargent JD. Effect of visual media use on school performance: a prospective study. J Adolescent Health. 2010;46(1):52–61. 29. Shin N. Exploring pathways from television viewing to academic achievement in school age children. J Genet Psychol. 2004;165(4): 367–82. 30. Stevens TA, To Y, Stevenson SJ, et al. The importance of physical activity and physical education in the prediction of academic achievement. J Sport Behav. 2008;31(4):368–88. 31. Stroth S, Kubesch S, Dieterle K, et al. Physical fitness, but not acute exercise modulates event-related potential indices for executive control in healthy adolescents. Brain Res. 2009;1269:114. EPIDEMIOLOGY 22. Mo¨Qle T, Kleimann M, Rehbein F, et al. Media use and school achievement–boys at risk? Brit J Dev Psychol. 2010;28(3):699–725. 23. Mountjoy M, Andersen LB, Armstrong N, et al. International Olympic Committee consensus statement on the health and fitness of young people through physical activity and sport. Br J Sports Med. 2011;45(11):839–48. 24. Munasib A, Bhattacharya S. Is the Fidiot’s box_ raising idiocy? Early and middle childhood television watching and child cognitive outcome. Econ Educ Rev. 2010;29(5):873–83. 25. Pate RR, Mitchell JA, Byun W, et al. Sedentary behaviour in youth. Br J Sports Med. 2011;45(11):906–13. 26. Reed JA, Einstein G, Hahn E, et al. Examining the impact of integrating physical activity on fluid intelligence and academic performance in an elementary school setting: a preliminary investigation. JPAH. 2010;7(3):343. 2104 Official Journal of the American College of Sports Medicine http://www.acsm-msse.org Copyright © 2013 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited. II INTERNAL CONSISTENCY AND STABILITY OF THE CANTAB NEUROPSYCHOLOGICAL TEST BATTERY IN CHILDREN by Heidi J. Syväoja, Tuija H. Tammelin, Timo Ahonen, Pekka Räsänen, Asko Tolvanen, Anna Kankaanpää & Marko T. Kantomaa. Submitted manuscript. Submitted manuscript. Internal consistency and stability of the CANTAB neuropsychological test battery in children Heidi J Syväoja1,2, Tuija H Tammelin1, Timo Ahonen2, Pekka Räsänen3, Asko Tolvanen2, Anna Kankaanpää1 & Marko T Kantomaa1,4. 1 LIKES – Research Center for Sport and Health Sciences, Jyväskylä, Finland, 2 University of Jyväskylä, Finland, 3 Niilo Mäki Institute, Jyväskylä, Finland, 4 Imperial College London, U.K. Abstract The Cambridge Neuropsychological Test Automated Battery (CANTAB) is a computer-assessed test battery widely use in different populations. The internal consistency and one-year stability of CANTAB tests were examined in school-aged children. Two hundred-thirty children (57% girls) from 5 schools in the Jyväskylä school district in Finland participated in the study in spring 2011. The children completed the following CANTAB tests: a) visual memory (Pattern Recognition Memory [PRM] and Spatial Recognition Memory [SRM]), b) executive function (Spatial Span [SSP], Stockings of Cambridge [SOC], and Intra-Extra Dimensional Set Shift [IED]), and c) attention (Reaction Time [RTI] and Rapid Visual Information Processing [RVP]). Seventy-four children participated in the follow-up measurements (64% girls) in spring 2012. Cronbach’s alpha reliability coefficient was used to estimate the internal consistency of the nonhampering test, and structural equation models were applied to examine the stability of these tests. The reliability and the stability could not be determined for IED or SSP because of the nature of these tests. The internal consistency was acceptable only in the RTI task. The one-year stability was moderate-togood for the PRM, RTI, and RVP. The SSP and IED showed a moderate correlation between the two measurement points. The SRM and the SOC tasks were not reliable or stable measures in this study population. For research purposes, we recommend using structural equation modeling to improve reliability. The results suggest that the reliability and the stability of computer-based test batteries should be confirmed in the target population before using them for clinical or research purposes. Keywords: CANTAB, internal consistency, reliability, stability, children INTRODUCTION Computer technology can make valuable contributions to neuropsychological assessments (Cernich, Brennana, Barker, & Bleiberg, 2007; Parsey & SchmitterEdgecombe, 2013). Neuropsychological test batteries have been used for one-off psychological assessments, for evaluating changes over time, and for evaluating the effects of interventions (Lowe & Rabbitt, 1998). The Cambridge Neuropsychological Test Automated Battery (CANTAB) is one of the oldest computer-based test batteries used to evaluate neurocognitive functioning, particularly in clinical trials research. CANTAB is a Windows-based program administered via a touch screen computer. CANTAB tests are mainly nonverbal and allow investigation of visual and spatial memory, executive function, working memory, planning, different aspects of attention, and other areas of cognition (Cambridge Cognition Ltd., 2006). Previous research suggests that CANTAB is a suitable method to measure cognitive functions in 4- to 90-year-old individuals, particularly in people with Alzheimer’s, Parkinson’s disease, schizophrenia, and autism (Lowe & Rabbitt, 1998; Luciana, 2003). Studies have also demonstrated that it is sensitive to deficits due to several neuropsychological and psychiatric conditions, especially in the elderly (e.g., Blackwell et al., 2003; Levaux et al., 2007; Robbins et al., 1998; Sahakian & Owen, 1992), but also in children (e.g., Gau & Shang, 2010; Fried, Hirshfeld-Becker, Petty, Batchelder, & Biederman, 2012; Luciana, Lindeke, Georgieff, Mills, & Nelson, 1999; Rhodes, Riby, Matthews, & Coghill, 2011). Although studies have shown that CANTAB tests can discriminate clinical populations from normal controls, little is known about the validity of these tests compared to traditional neuropsychological tests in the general population (Smith, Need, Cirulli, Chiba-Falek, & Attix, 2013). CANTAB tests are mainly based on traditional neuropsychological tests (Cambridge Cognition Ltd., 2006). These traditional tests have been highly used and their validity and reliability has been carefully assessed (Ahonniska, Ahonen, Aro, Tolvanen& Lyytinen, 2000; Gnys & Willis, 1991; Mammarella, Pazzaglia, & Cornoldi, 2008; Halperin, Sharma, Greenblatt, & Schwartz, 1991). However, the reliability of the CANTAB tests has been inadequately described in earlier studies (Luciana & Nelson, 2002). According to Luciana (2003), internal consistency coefficients were high (.73 – .95) in 4–12-year-old children. However, to our knowledge, there are no other studies establishing the internal consistency agreement of CANTAB tests in a child population. Furthermore, studies measuring the testretest reliability of CANTAB have been sparse. According to Lowe and Rabbit (1998), the test-retest agreement for the CANTAB tests in an elderly adult population was either moderate, ranging from .70 to .86, or low, ranging from .09 to .68 for four week time interval. Fisher et al. (2011) observed quite low intra-class correlations for CANTAB subtests, Spatial Span length, and Spatial Working Memory errors (ICC = .51 – .59) for a three-week interval in healthy children. However, according to Gau and Shang (2010), intra-class correlations for CANTAB tests (Intra-Extra Dimensional Set Shift, Spatial Span, Spatial Working Memory, Stockings of Cambridge) ranged from .55 to .94 for time interval of 14–42 INTERNAL CONSISTENCY AND STABILITY OF THE CANTAB days in a group of 10 children with Attention deficit hyperactivity disorder (ADHD). To our knowledge, there are no other studies establishing the test-retest agreement for CANTAB tests in a child population (Henry & Bettenay, 2010; Luciana, 2003), and the results of existing studies are, to some extent, inconsistent. It is important to examine the reliability and the stability of the tests because low reliability limits both the sensitivity to diagnose clinical conditions and the sensitivity to detect changes in cognition over time (Lowe & Rabbitt, 1998). In addition, low reliability and stability limit the usefulness of the test as a research and clinical tool. Moreover, the use of CANTAB and similar computer-based neuropsychological test batteries is increasing worldwide and that is why it is important to determine the psychometric properties of these kinds of test batteries. The purpose of this study was to evaluate the internal consistency and the one-year stability of seven CANTAB tests measuring visual memory, executive function, and attention in elementary school-aged children. 3 letter for their parents/guardians, and a consent form. Participation in the study was voluntary, and all the participants and their parents were informed about their right to drop out of the study at any time without a specific reason. Only children with a fully completed consent form (signed by a parent/guardian and the child) on the day of the first measurements were included in the study. In 46% of families, the highest level of parental education was tertiary level education. Seventy-seven percent of the parents were married or cohabiting. Six percent of the children had a diagnosed learning difficulty. The children had normal or corrected-to-normal vision. They participated in normal curriculum-based instruction, and the language of instruction was Finnish. During spring 2012, students who were fifth graders in spring 2011 were invited to participate in the follow-up measurements. Seventy-four children participated in these follow-up measurements (49% of 151 eligible, 64% girls, Mage2011 = 11.73, SD = .37, Mage2012 = 12.82, SD = .04). The study was performed according to the principles of the Declaration of Helsinki and the Finnish legislation and was approved by the Ethics Committee of the University of Jyväskylä. METHOD Participants The data of the present study is part of a larger research project, which aims to determine the associations of physical activity, sedentary behavior, cognitive functions and academic achievement in elementary school-aged children. In spring 2011, 230 fifth and sixth graders (48% of 475 eligible, 57% girls, Mage = 12.19, SD = .63) from five schools in the Jyväskylä school district in Finland participated in the study. When the children were invited to participate in the study, they were given an information pack containing a leaflet for themselves, a Procedures CANTAB (a PaceBlade Slimbook P110 tablet PC with a 12-inch touch-screen monitor and Windows XP Professional operating system, CANTABeclipse version 3) was used to assess a broad range of cognitive functions: a) visual memory (Pattern Recognition Memory [PRM] and Spatial Recognition Memory [SRM]), b) executive function (Spatial Span [SSP], Stockings of Cambridge [SOC], Intra-Extra Dimensional Set Shift [IED]), and c) attention (Reaction Time [RTI] and Rapid Visual Information Processing [RVP]) (Table 1). The tests were run individually with the INTERNAL CONSISTENCY AND STABILITY OF THE CANTAB help of a trained research assistant and according to the standard protocol. The standard instructions for the tests were provided in the CANTAB manual and were translated into Finnish. The execution of the tasks required about 45 minutes. The test battery was administered in a silent room without distractions. A Motor Screening Task measuring simple psychomotor speed and accuracy was used as a training procedure at the beginning of a test session. Measures Visual memory. Visual memory performance was assessed with PRM and SRM. PRM measures recognition memory for visual patterns in a two-choice forced discrimination paradigm. In the presentation phase, the children were presented with 12 different geometric patterns one after the other in the center of the screen. Beforehand, they were asked to remember the patterns. In the recognition phase, the children were presented with 12 trials of two patterns: a pattern they had already seen and a novel pattern. They were asked to choose a pattern they remembered having seen. The target patterns were presented in the same order as in the first time. This subtest was repeated with a new set of the 12 patterns to be remembered. The score in this task is based on the number of correct responses (maximum 24). SRM measures recognition memory for spatial locations in a forced-choice paradigm. In the presentation phase, the children were presented with a white square on the screen in five different locations and asked to remember the locations where they had seen the square. In the recognition phase, the children were shown two squares in different locations: one in the same location as before and the other in a new location. The children were instructed to choose the location where they remembered seeing the square. The 4 target locations were presented in the same order as in the first time. The block of five trials was repeated four times in total. The score in this task is based on the number of correct responses (maximum 20). Executive function. Children’s executive functions were assessed with SSP, SOC and IED tests. The SSP is based on the Corsi Blocks task (Milner, 1971), which measures the length of the visuospatial memory span. In each trial, there are 10 white boxes on the screen, and the color of a specified number of boxes changes one by one. The children were directed to reproduce the sequence by touching the same boxes in the same order that the boxes changed their color. If the child reproduced the correct sequence, he/she passed to the next difficulty level, where one more box was added to the sequence. The child has three attempts at each level. If the third attempt was unsuccessful, the task was terminated. The task starts with a two-box sequence and ends with a nine-box sequence, which is the highest possible level to proceed. The score in the task is based on the length of the maximum sequence that the child can reproduce. The SOC is a computerized version of the Tower of London task (Owen, Downes, Sahakian, Polkey, & Robbins, 1990; Shallice & Shallice, 1982) measuring spatial planning and spatial working memory. At the beginning of each problem, the children were presented with a computer screen split into two parts. In both parts of the screen, there were three vertical stockings and three colored balls in predetermined order. The children were required to move the colored balls in the lower part of the screen to the same position in the stockings as they are in the upper part of the screen. They were asked to use only a specific number of moves (two, three, four, or five) to fulfill the goal. The balls can be moved one at time by touching INTERNAL CONSISTENCY AND STABILITY OF THE CANTAB first the required ball and then the target location. The balls cannot be moved outside the stockings, and the lower balls cannot be moved before the upper balls. If the child took more than double the number of moves required to fulfill the goal, the task was terminated. If three consecutive problems were terminated, the entire test ended. The score in this task is based on the number of problems the child solves with the minimum number of moves. IED is a computerized analogue of the Wisconsin Card Sorting test and measures rule acquisition and reversal in a set-sifting condition. Specifically, it measures the ability to focus attention on different stimuli within a relevant dimension and shift attention to a previously irrelevant dimension. There are nine stages, with increasing difficulty in this task. The children were instructed to choose one of the two different dimensions: one is correct, and the other is incorrect. According to immediate feedback given from the computer, they were expected to choose the correct pattern and learn the rule. The children progressed through the task by satisfying a predetermined criterion of learning the rule at each stage (six consecutive correct choices). If the children failed to learn the rule during 50 trials at any stage, the test was terminated. The first two stages of the task measured simple discrimination (stage 1) and simple reversal (stage 2), and the children had to choose between two purple patterns. In stages 3 and 4, compound stimuli (a purple pattern and a white lined drawing) were presented. In these stages, the children had to continue to respond to the previously relevant dimension (the purple pattern) and ignore the presence of the new irrelevant dimension (the white lined drawing) (nonoverlapping compound discrimination [stage 3] and compound overlapping discrimination [stage 4]). These 5 stages were followed by compound reversal (stage 5). In stage 6, the first attentional shift is required (the intradimensional shift). The children were presented with new compound stimuli: novel shapes of each of the two dimensions (the purple pattern and the white line drawing). They had to continue to respond to the relevant dimension (the purple pattern). This stage was followed by the intradimensional reversal (stage 7). In stage 8, the compound stimuli changed again, but this time the children had to shift their attention (the extradimensional shift) and respond to the previously irrelevant dimension (the white lined drawing). Stage 9 involves extradimensional reversal. The score in this task is based on the number of stages completed. Attention. The children’s attention abilities were assessed with RTI and RVP. RTI measures the children’s speed of response to an unpredictable visual target. In the unpredictable condition, the yellow spot appears in any of five circles on the screen. The children were instructed to hold down the press pad button until they saw the yellow spot and then touch the middle of the correct circle as quickly as possible. The children took part in rehearsal trials and 15 task trials. The scores in this task are based on reaction time (ms) and movement time (ms). The RVP measures the sustained attention and is similar to the Continuous Performance Task. There is a white box on the screen where digits from 2 to 9 appear in a pseudorandom order at the rate of 100 digits per minute. The children were directed to touch the press pad button every time they saw the following target sequence: digits 3, 5, and 7 (the 357 mode). During the practice stage, the children received hints and feedback from the computer, which declined gradually. In the assessment stage, the children received no hints or feedback. This stage took 3 minutes. INTERNAL CONSISTENCY AND STABILITY OF THE CANTAB The score in this task is based on RVP A’, which measures how good the child is at detecting the target sequences (range: .00 to 1.00; bad to good). Statistical analysis For the statistical analyses, the SPSS 19.0 for Windows statistical package (SPSS (2010) IBM SPSS Statistics 19 Core System User’s Guide (SPSS Inc., Chicago, IL) and the Mplus statistical package (Version 7; Muthèn & Muthèn, 1998–2012) were used. Logarithmic transformations were applied to variables with skewed distributions, and gender differences were tested using the independent samples t-test. The differences between the subjects with complete data and the subjects who did not participate in the follow-up measurement were tested using the independent samples t-test for continuous variables and Pearson’s Chi-Square test for dichotomously scored variables. Pearson’s correlation coefficients were calculated for continuous variables between the different sections of each test. Effect sizes (Cohen’s d, standardized mean difference) were calculated for repeated measures. The reliability of each nonhampering test was estimated with Cronbach’s alpha reliability coefficient (α). As preliminary analysis of stability, Pearson’s correlation for continuous variables and tetrachoric correlations for dichotomously scored variables were calculated between the assessments. To examine the stability of the CANTAB tests, structural equation models were applied. To combine a very large number of measured variables for each latent factor, item parcels were constructed by summing every third pattern to same parcel (Little, Cunningham, Shahar, & Widaman, 2002). Item parcels were used in order to achieve continuous indicators and not to end up with too large model in terms of the ratio of sample size to number of free parameters (Herzog & 6 Boomsma, 2009; Westland, 2010). Underlying latent traits were assumed to be unidimensional. The constructed parcels were then used as indicator variables in the confirmatory factor analyses. The measurement models were first specified at both measurement times to test the association between the observed variables and the underlying factors. After demonstrating the fit of the measurement models, longitudinal confirmatory factor analyses were performed. The baseline stability model, in which the factor (factors) in the second assessment (2012) was predicted by the factor (factors) in the previous measurement point (2011), was estimated. To detect time invariance in the latent constructs, equality constraints were imposed on the corresponding factor loadings across two time points. Furthermore, if the invariance assumption of a stability model was supported, a more parsimonious model in which all the factor loadings were fixed to be one was estimated. Full information maximum likelihood (FIML) estimation with robust standard errors (MLR) was used under the assumption of data missing at random. Item response theory (IRT) modeling using FIML estimation was applied to the CANTAB tests with hampering nature. The goodness-of-fit of the cross-sectional and longitudinal models was evaluated by the Satorra–Bentler scaled χ2-test, the comparative fit index (CFI), the Tucker– Lewis Index (TLI), the root mean square error of approximation (RMSEA), and the standardized root-mean-square residual (SRMR). The model fits the data well if the p-value for the χ2-test is non-significant. CFI and TLI values close to 0.95, an RMSEA value below 0.06, and an SRMR value below 0.08 indicate good fit between the model and INTERNAL CONSISTENCY AND STABILITY OF THE CANTAB the observed data (Hu & Bentler, 1999). A Satorra–Bentler scaled χ2 difference test was conducted for the nested models. If the χ2-test produces a non-significant loss of fit for the constrained model as compared to the baseline stability model, the equality constraints are supported. RESULTS In this study, Cronbach’s alpha reliability coefficients were calculated for crosssectional data in 2011 to estimate the internal consistency of the CANTAB tests. Structural equation modeling was used to estimate the one-year stability of the CANTAB tests in longitudinal data setting. The mean values, standard deviations, and gender differences for the cross-sectional sample at the first measurement point are presented in Table 2. The performance of the boys and girls did not differ in the CANTAB test, except in the RTI, where the boys’ fivechoice movement time (2011: t(228) = 3.30, p = .001, 2012: t(86) = 2.81, p = .006) and reaction time (2011: t(228) = 4.07, p < .001, 2012: t(86) = 2.39, p = .020) were shorter than that of the girls. The mean values and the standard deviations for the follow-up sample and the effects sizes (Cohen’s d) for the repeated measures are presented in Table 3. The subjects with complete data (n = 74) did not differ from the subjects who did not participate in the follow-up measurements (n = 156) with respect to the highest level of parental education, family structure, family income, or diagnosed learning difficulties. These groups also did not differ in their performance in the CANTAB tests. Reliability Cronbach’s alpha reliability coefficients were .65 for the PRM number of correct responses, .21 for the SRM number of correct responses, 7 .87 for the RTI five-choice movement time, .66 for the RTI five-choice reaction time and .49 for the RVP A’. For hampering tests (SSP, SOC and IED), Cronbach’s alpha could not be determined. Stability Visual memory. The distribution of the PRM number of correct responses was negatively skewed, and 50% of children made two mistakes or less. For the PRM number of correct responses, Pearson’s correlation between the assessments was r(74) = .53 (p < .001 ). Three parcels from 24 patterns (incorrect/correct response) of PRM were constructed by summing eight items into the same parcel. These subscales were then used as the indicator variables in the confirmatory factor analyses. The baseline stability model fitted the data well. Invariance assumption was supported, with an insignificant scaled difference in the χ2 value (χ2 (2) = .89, p = .64). In addition, the model in which all the loadings were fixed to be one was confirmed (χ2 (4) = 4.67, p = .32). The goodness-of-fit statistics of the more constrained model for the PRM number of correct responses were good (χ2 (12) = 17.30, p = .14, CFI = .96, TLI = .95, RMSEA = .04, SRMR = .13). The estimation results of the model are presented in Figure 1. The stability for the PRM was .80. The factor loadings and the measurement error variances were significant. Three parcels from 20 patterns (incorrect/correct response) were formed also for the SRM. The estimation results of the cross-sectional measurement models in 2011 and 2012 revealed that none of the parcels loaded significantly on the hypothesized factor. Therefore, a stability model for SRM could not be determined. Correlations among 20 trials of SRM were calculated, and only 25 correlations of 190 were statistically significant. The correlations were modest and ranged from -.14 to .37, with one exception: INTERNAL CONSISTENCY AND STABILITY OF THE CANTAB The second and fifth trials in the third block correlated with each other (r(229) = .77, p < .001). The correlation for the SRM number of correct responses between 2011 and 2012 was r(72) = .30 (p = .009). The level of performance remained the same in 20% of the children, it improved in 43%, and it declined in 37%. Executive function. The stability model for the SSP could not be determined, because of the nature of the test. The distribution of children’s span lengths are presented in Table 4. The correlation between the SSP span length in 2011 and in 2012 was r(72) = .37 (p = .001). The span length stayed the same in 34% of children, improved in 53%, and declined in 14%. The cross-sectional IRT models for SOC were estimated for dichotomously scored items (problem solved/not solved in minimum moves). Only 5 of the 12 items loaded significantly on the hypothesized factor in 2011 and 3 of the 12 items in 2012. According to the results of the estimated baseline stability model, there was no significant association between the latent variables measured in 2011 and 2012. The correlation for SOC problems solved in the minimum number of moves between 2011 and 2012 was r(72) = .23 (p = .046). The level of performance stayed the same in 11% of children, improved in 58%, and declined in 31%. As the numbers of errors in stages 3 to 7 and 8 to 9 of the IED test were highly correlated with each other, a cross-sectional two-factor model of the number of errors was estimated. All the items loaded on the hypothesized factors, except for the errors in stage 7 in 2012. There was no significant correlation between the factors either in 2011 or in 2012. In addition, the results of the estimated baseline stability model revealed that the 8 regression coefficient between the latent variables measured in consecutive years was significant only for the latent variable measured by the errors in stages 8 and 9 (b = .68, SE = .08, p < .001). Therefore, the stages from 3 to 7 were discarded from further analyses. For the dichotomous variables, which indicated whether the child completed the stage, tetrachoric correlations were calculated. The tetrachoric correlations between stages 8 and 9 were .99 (p < .001) in 2011 and .96 (p < .001) in 2012. The tetrachoric correlation for stage 8 between 2011 and 2012 was .38 (p = .047). For stage 9, it was .64 (p < .001). The level of performance remained the same in 69% of the children, it improved in 19%, and it declined in 12%. In 2011, 66% of the children passed the test, and 72% passed the test in 2012. Attention. For the RTI five-choice reaction time, Pearson’s correlation between the assessments was r(74) = .63 (p < .001). For the RTI five-choice movement time, it was r(74) = .50 (p < .001). For structural equation modeling, three subscales of reaction time and movement time at both measurement points were formed from 15 patterns in the RTI (and each subscale was divided by 10). The results of the estimated cross-sectional two-factor model confirmed that the subscales of reaction time and movement time loaded on their hypothesized factors. The baseline stability model fitted the data reasonably well. The scaled χ2-difference test produced a non-significant loss of fit for the constrained stability model compared with the baseline model (χ2 (4) = 6.59, p = .16). In addition, a more constrained model, in which the loadings were fixed to be one, was supported, with an insignificant difference in the χ2 value (χ2 (8) = 6.91, p = .55). The more constrained stability model for the RTI reaction time and the movement time is presented in Figure 2. The goodness-of-fit INTERNAL CONSISTENCY AND STABILITY OF THE CANTAB statistics of the more constrained model for the RTI reaction time and the movement time were good (χ2 (58) = 78.51 (58) p = .04, CFI = .96, TLI = .95, RMSEA = .04, SRMR = .15). The stability coefficient for the RTI movement time was .67. For the RTI reaction time, it was .78. All the factor loadings and the measurement error variances were significant. For the RVPA’, Pearson’s correlation between the assessments was r(74) = .39 (p = .001). The RVP A’ (multiplied by 10) scores of the three blocks of the test were used as indicator variables in the structural equation modeling. The distribution of the RVP A’ number of correct responses was highly negatively skewed. The baseline stability model fitted the data reasonably well. The invariance assumption was supported, with an insignificant difference in the χ2 value (χ2 (2) = 4.41, p = .11). When all the factor loadings were constrained to be one, a scaled χ2 difference test produced a significant loss of fit (χ2 (4) = 10.96, p = .03). A constrained stability model for RVP A’ is presented in Figure 3. The goodness-of-fit statistics of the constrained model for RVP A’ were reasonable good (χ2 (10) = 17.94 (10) p = .06, CFI = .91, TLI = .87, RMSEA = .06, SRMR = .25. The stability coefficient for RVP A’ was .62. All the factor loadings and the measurement error variances were significant. DISCUSSION In this study, the internal consistency and the one-year stability of seven CANTAB tests were examined in elementary school-aged children. According to Cronbach’s alpha reliability coefficients, only the RTI fivechoice movement time increased above a typically accepted level of .7 (Nunnally & Bernstein, 1994). According to the structural 9 equation modeling, the PRM number of correct responses and the RTI reaction time had high levels of stability. In addition, the RTI movement time and the RVP A’ had a moderate level of stability. Furthermore, the SSP span length and the IED number of children who completed at least stage 8 seemed to have a moderate correlation between the two measurement points. The SRM and the SOC tasks were not reliable or stable measures in this study population. Visual memory. Cronbach’s alpha reliability coefficients for the PRM and the SRM were below the accepted level of .7 (Nunnally & Bernstein, 1994), indicating that the total scores of the tests should be used with caution. In addition, the patterns in the SMR did not correlate with each other, and those patterns that did showed only a modest correlation, fluctuating around zero. This finding does not support that of Luciana (2003) who reported that the internal consistency coefficients for CANTAB tests ranged from .73 to .95 in 4–12-year-old children. To increase the reliability of the PRM, we recommend estimating the measurement errors away by using structural equation modeling in the analyses. One possible explanation for the low internal consistency of the PRM may be the ceiling effect. According to Luciana and Nelson (2002), children reach an adult level of performance in the PRM by the age of 7 years, after which ceiling levels are reached. Thus, it may be problematic to discriminate 9- to 12-year-old children with high ability. In our study, the children were about 12 years. Generally, their performance was high in the PRM test. Fifty percent of the children made two errors or less in the test. It seems that the errors the children made were sparse and random, which may explain why the patterns of the tests do not have a high correlation with each other. However, this INTERNAL CONSISTENCY AND STABILITY OF THE CANTAB was not the case with the SRM because the children’s performance did not reach the level of ceiling. The stability was good for the PRM number of correct responses (.80). However, a stability model for SRM could not be determined because none of the parcels of SRM loaded significantly on the hypothesized factor. The finding regarding PRM’s stability is consistent with that of Lowe and Rabbitt (1998) who reported that the test-retest correlation for PRM was .84 in a healthy adult population (ages 60–82 years). In summary, according to the results of this study, the PRM task of CANTAB proved to be a stable measure for 12-year-olds. To increase the reliability of the PRM, we recommend using structural equation modeling in research analyses. The SRM task was not a reliable measure. Executive function. The traditional version of Corsi Blocks test has been shown to be a reliable test for young adults (α = .85) (Miyake, Friedman, Rettinger, Shah, & Hegarty, 2001) and children (α = .79) (Mammarella et al., 2008). In this study, the CANTAB version of the Corsi blocks task, SSP, seemed to be an adequate tool for measuring visuospatial memory capacity in the children aged 12 years. There was variance in the results of the SSP among the children. In addition, most of the children improved their performance one year later, which is line with a previous study that found that children approximates functional maturity in the SSP task at the age of 12 years (Luciana & Nelson, 2002). Nevertheless, we cannot draw advanced conclusions about the reliability of the SSP because its internal consistency could not be determined. Thus, we cannot say how much is real variance and how much is measurement error. 10 In the case of the SSP, either stability model could be determined. However, the correlation for the SSP span length between 2011 and 2012 was .37. According to previous studies, the test-retest reliability for SSP was .51 for a three-week interval (Fisher et al., 2011), .55 for 14–42 days interval (Gau & Shang, 2010) in children and .64 for a fourweek interval in the elderly (Lowe & Rabbitt, 1998). The CANTAB test, SOC is identical to the traditional Tower of London (TOL) task (Owen et al., 1990; Shallice & Shallice, 1982). According to the results of this study, it seems that the patterns in the SOC are not informative. Other studies had also raised questions about the psychometric characteristics of traditional TOL tasks (Bishop, Aamodt-Leeper, Creswell, McGurk, & Skuse, 2001). Humes, Welsh, Retzlaff, and Cookson (1997) reported low internal consistency for the TOL (split-half reliability of .19 and Cronbach alpha of .25). In addition, according to Ahonniska et al. (2000), a similar task, the Tower of Hanoi (TOH) did not have satisfying reliability (simplex estimator) in the first two assessments, but improved with repetition, when children aged 8 and 12 years participated in nine repeated assessments of the task during 18 months. The test-retest reliability of the traditional TOL task has been reported to be quite low: .5 for 30–40 day time interval (Bishop et al., 2001), but also acceptable: .72 for 25 min time interval in preschool children (Gnys & Willis, 1991). However, Ahonniska et al. (2000) reported relatively high stability for TOH (Beta= .82– 1.00, depending on the performance index) after the second assessments. The temporal stability of the CANTAB version, SOC, has been reported to quite low: 26–.60 for four week time interval (depending on the performance index) (Lowe & Rabbitt, 1998), but also acceptable: .72 for 14–42 day time INTERNAL CONSISTENCY AND STABILITY OF THE CANTAB interval in children with ADHD (Gau & Shang, 2010). Ahonniska et al. (2000) proposed that large intraindividual variation due to different rates of learning may explain the low reliability (simplex estimator) in the first few assessments. Also, Lowe and Rabbitt (1998) suggested that one possible explanation for the low temporal stability of SOC might be the novelty of the task. Performance in the tests of executive function can abruptly improve when an individual discovers an optimal strategy, but the performance improves less or not at all if a strategy is not found. The performance may even decline if an incorrect strategy is attempted. Different practice effects will weaken the test-retest reliability (Lowe & Rabbitt, 1998). According to Anderson, Anderson, and Lajoie (1996), performance in the TOL improves approximately at the age of 11 years due to a developmental spurt. In our study, 58% showed improved performance, and 31% showed a decline in performance. However, Bishop et al. (2001) reported that average scores in the retest and the initial test were nearly the same in 7–15-year-old children indicating that task novelty cannot account for the low test-retest reliability. Bishop et al. (2001) suggested that any variation due to individual differences in neurology in executive function performance may be overwhelmed by powerful factors other than brain development influencing the performance in executive function tasks. In addition, it has been reported that a simpler TOL (two, three, four, or five moves) gives ceiling effects in older children (Anderson et al., 1996; Krikorian, Bartok, & Gay, 1994). However, according to Luciana, Collins, Olson, and Schissel (2009) and Luciana and Nelson (2002), 11–12-year-old children do not reach adult levels of performance. In our study, the children did not reach the ceiling 11 levels, which suggests that the ceiling effect does not explain the low reliability in this case. IED task is based on traditional the Wisconsin Card Sorting test. In this study, the analysis showed that only stages 8 and 9 were informative, and around 70% of the children passed the whole test. According to previous studies, shifting ability improves with age and reaches ceiling levels by age 12 years (Anderson, 2002; Luciana & Nelson, 2002). According to Luciana and Nelson (1998), it is typical for normal adults to make significantly more errors in the extradimensional shift stage (at stage 8) than in earlier stages. This phenomenon was repeated in the 8-year-olds but not in the younger children (Luciana & Nelson, 1998). Substantial amount of errors in the extradimensional shift stage was also observed in the current study with 12-yearolds, and it seems that only stages 8 and 9 discriminate children with weaker performance from children with general performance. In this study, according to the tetrachoric correlations, the minimum stability for IED stage 8 completed was 0.38. For IED stage 9 completed, it was .64. In previous studies, the temporal stability for the CANTAB IED task was reported to be .78 for 14–42 day time interval (Gau & Shang, 2010) in children, .40–.75 (depending on the performance index) in adults (Henry & Bettenay, 2010) and .09–0.70 (depending on the performance index) for four week time interval in the elderly (Lowe & Rabbitt, 1998). In summary, it seems that only the SSP of these executive function tests could work well in healthy 12-year-old children. The IED was quite stable, but it seemed to be too easy for the 12-year-olds and did not discriminate between children with higher abilities. In this INTERNAL CONSISTENCY AND STABILITY OF THE CANTAB study, the SOC was not a reliable or stable measure. Attention. In this study, the internal consistency coefficient for the RTI fivechoice movement time reached the accepted level of .7 (Nunnally & Bernstein, 1994), whereas internal consistency was below this accepted level for the RTI five-choice and RVPA’. The stability was good for the reaction time and moderate for the RTI movement time and the RVP A’. Sustained attention has been assessed with the Continuous Performance Task before, which is similar to the CANTAB RVP task. The split-half reliability for the computerized Continuous Performance Task was reported to be .38–.95 (depending on the performance index) (Conners, Epstein, Angold, & Klaric, 2003; Halperin et al., 1991). According to previous studies, the test-retest reliability ranged from .55–84 (depending on the performance index) for 4.8 month interval (Conners et al., 2003) and 0.55–.84 (depending on the performance index) for three month interval (Halperin et al., 1991). In this study, the RVP A’ performance of the children was very high. According to Halperin et al. (1991), there were no difficulties related to ceiling or floor effects in the Continuous Performance Test in 7–10year-old boys, but they assumed that ceiling effects would occur in older children. According to the results of this study, the ceiling effect was observed in the RVP task, which may explain the low internal consistency. In addition, the low stability may also be the result of the easiness of the test. Due to high level of performance, a single random mistake may induce the effect of varying performance between the assessments, and affect stability. In summary, according to the results of this study, the RTI task of CANTAB is a reliable 12 and stable measure for 12-year-olds. When using this test in research analysis, we recommend estimating the measurement errors away by using structural equation modeling to improve the reliability of the RTI. The RVP 357 mode was easy for the 12year-olds in this study. It might be useful to utilize a more difficult version of the test with a greater number of target sequences in this age group. Taken together, the internal consistency of most of the CANTAB tests used in this study was low, and two of the tests did not measure the phenomena they were supposed to measure with satisfying reliability. Some previous studies reported that computerized versions of traditional neuropsychological tests are not equivalent to traditional manual tests. For example, Feldstein et al. (1999) reported that computerized versions of the Wisconsin Card Sorting test were not similar to the manual version. In addition, according to Smith et al. (2013), CANTAB subtests measuring executive function, speed of processing, visual memory, and working memory correlate only modestly with traditional subtests. The reason for these findings is unknown. Perhaps, computerized test sessions are more sensitive to attentional disruptions than manual sessions, or they may not offer similar perceptual characteristics to manual versions. In the present study, stability of the measured CANTAB tests was moderate supporting the previous studies (Fisher et al., 2011; Gau & Shang, 2010; Lowe & Rabbitt, 1998). However, in these previous studies the testretest interval has been shorter, only few weeks, whereas, in the present study, stability was measured with one year interval. This longer time interval may affect the stability, because it is expected that the cognitive abilities of 11-year-old children develop INTERNAL CONSISTENCY AND STABILITY OF THE CANTAB during one year. Therefore, measuring their performance twice in the beginning with a shorter interval between the measures would have enabled calculation of the test-retest reliability, and ruling out the developmental effects. Lastly, there are a lot of advantages in computerized testing: computer technology has increased the efficiency, ease, and standardization of administration and saved time and money related to testing. Electronic data capture and automatic results scoring have minimized human errors in scoring and data entry and increased the accuracy of timing and response latencies. (Cernich et al., 2007; Parsey & Schmitter-Edgecombe, 2013). Other advantages of computerized technology – particularly for the measurement of attention, motor, and memory functioning – are the availability of almost unlimited alternate forms and an increased number of trials. This minimizes practice effects and allows more assessments at shorter time intervals compared to traditional measures. A large number of trials and accurate assessment of reaction times results in data that is normally distributed and on a true interval scale (Betts, Mckay, Maruff, & Anderson, 2006). This is particularly important when slight changes in performance across time are assessed. Computer technology also provides test administration conditions that are accommodating for individuals with particular needs (American Educational Research Association [AERA], American Psychological Association [APA] & National Council on Measurement in Education [NCME], 2014). Touch-screen technology facilitates use by young children and certain clinical groups, and allows more reliable assessment of motor function and processing speed compared to traditional measures using an individual administrator wielding a 13 stopwatch. In addition, non-verbal cultureneutral test stimuli are often used, which allow the application of computerized test batteries (like CANTAB) for individuals from different racial, ethnic, geographic, or sociocultural backgrounds. (Luciana, 2003; Henry, 2010). CANTAB also has simple standardized test administration; thus, it is easy to use and no previous IT or scientific training is needed to set-up and administer the test (Cambridge Cognition Ltd., 2006). Despite the potential advantages that computerized technology offers for neuropsychological testing – especially the ease of building ready algorithms for calculating indexes and scores – computerassisted assessment also produces a risk factor. Due to commercial competitive reasons, the companies providing these tools may not always publish detailed information about these algorithms or the psychometric properties of the tasks. The general characteristics of scoring algorithms and the accuracy of the algorithms as well as technical evidence should be documented and reviewed periodically (AERA, APA & NCME, 2014). According to The Finnish Psychological Test Committee (2013), test batteries (computerized or not) that do not provide a satisfactory level of information about the psychometric properties of the tasks in the technical manual should not be used in clinical practice. Likewise, in clinical practice, an expert (a psychologist or an MD) should always do the interpretation of the test results. There are two reasons for this. First, to use CANTAB or other computerized test batteries, there needs to be a person who supervises and paces the introduction of tests and monitors the assessment process (Luciana, 2003; AERA, APA & NCME, 2014). Secondly, the interpretation and the clinical conclusions and decisions made based partly on the test results produce a juridical situation where there needs to be a INTERNAL CONSISTENCY AND STABILITY OF THE CANTAB responsible decision-maker. Most countries do not allow automated decision-making in healthcare. There will be rapid growth in the usage of computer-assisted test batteries. However, little is known about the psychometric properties of computerized batteries as well as how performance on computerized batteries correlates with traditional neuropsychological measures. In addition, it is also vital to have more information about the reliability and validity of these batteries in all clinical target populations of interest as well as in samples derived from the normal population. Strengths and limitations To our knowledge, this is the first study examining the internal consistency and oneyear stability of the CANTAB tests in healthy 12-year-old children. This study provides valuable and important information on the psychometric characteristics of CANTAB tests in a child population. The study sample in this study in 2011 was quite large and representative of Finnish children at the age of 12. On the other hand, the number of children in the follow-up measurements in 2012 was limited, which attenuated the statistical power. However, subjects with complete data did not differ from the subjects who did not participate to follow-up measurements in socioeconomic positions or performance in the CANTAB tests. In addition, the models were estimated with FIML estimation, which uses all information available and takes missing data into account. The age-range of the children studied was narrow, which limits the application of the results to different age groups or developmental stages. In addition, the study sample was culturally homogeneous including children only from Finland. Also, we did not measure the performance of the children again immediately after the first 14 measurement. Thus, we could not calculate intra-class correlations for the CANTAB tests. Furthermore, several measurement points during the year would give more accurate information on the effects of practice on the performance in the tests. Summary and conclusion In this study, psychometric characteristics of seven CANTAB tests were determined. According to the results, the internal consistency was acceptable only in the RTI task. The one-year stability was moderate-togood for the PRM, SSP, IED, RTI, and RVP. The SRM and SOC tasks were not reliable or stable measures in this study. In addition, the PRM, IED, and RVP had ceiling effects in the present study population of 12-year-old healthy children. The results of this study also suggest that psychometric characteristics of traditional neuropsychological tests may not remain when transplanted into computer form. Thus, the reliability and the stability of CANTAB and other computer-based test batteries among the target population should be confirmed before using them for clinical or research purposes. Acknowledgments This study was funded by Finnish Ministry of Education and Culture and the Academy of Finland (grant 273971). REFERENCES Ahonniska, J., Ahonen, T., Aro, T., Tolvanen, A., & Lyytinen, H. (2000). Repeated assessment of the Tower of Hanoi test: Reliability and age effects. Assessment, 7(3), 297–310. doi: 10.1177/107319110000700308 American Educational Research Association, American Psychological Association, National Council on Measurement in Education. (2014). INTERNAL CONSISTENCY AND STABILITY OF THE CANTAB Standards for educational and psychological testing. Washington DC: American Educational Research Association. Anderson, P. (2002). Assessment and development of executive function (EF) during childhood. Child Neuropsychology, 8(2), 71–82. doi: 10.1076/chin.8.2.71.8724 Anderson, P., Anderson, V., & Lajoie, G. (1996). The Tower of London test: Validation and standardization for pediatric populatons. The Clinical Neuropsychologist, 10(1), 54–65. doi: 10.1080/13854049608406663 Betts, J., Mckay, J., Maruff, P., & Anderson, V. (2006). The development of sustained attention in children: The effect of age and task load. Child Neuropsychology, 12(3), 205-221. doi: 10.1080/09297040500488522 Blackwell, A. D., Sahakian, B. J., Vesey, R., Semple, J. M., Robbins, T. W., & Hodges, J. R. (2003). Detecting dementia: novel neuropsychological markers of preclinical Alzheimer’s disease. Dementia and geriatric cognitive disorders, 17(1-2), 42–48. doi: 10.1159/000074081 Bishop, D., Aamodt‐Leeper, G., Creswell, C., McGurk, R., & Skuse, D. (2001). Individual differences in cognitive planning on the Tower of Hanoi task: Neuropsychological maturity or measurement error? Journal of Child Psychology and Psychiatry, 42(4), 551– 556. doi: 10.1111/1469-7610.00749 Cambridge Cognition Ltd. (2006). CANTABeclipsetm: Software User Guide. Manual version 3. Cambridge: Cambridge Cognition Ltd. (2006). Cernich, A. N., Brennana, D. M., Barker, L. M., & Bleiberg, J. (2007). Sources of error in computerized neuropsychological assessment. Archives of Clinical Neuropsychology, 15 22(Suppl 1), S39–S48. doi: 10.1016/j.acn.2006.10.004 Conners, C. K., Epstein, J. N., Angold, A., & Klaric, J. (2003). Continuous performance test performance in a normative epidemiological sample. Journal of Abnormal Child Psychology, 31(5), 555–562. doi: 10.1023/A:1025457300409 Feldstein, S. N., Keller, F. R., Portman, R. E., Durham, R. L., Klebe, K. J., & Davis, H. P. (1999). A comparison of computerized and standard versions of the Wisconsin card sorting test. The Clinical Neuropsychologist, 13(3), 303– 313. doi: 10.1076/clin.13.3.303.1744 The Finnish Psychological Test Committee (2013). Guidelines for using computerassisted tests in psychological assessment. [in Finnish]. Available at http://www.psyli.fi/files/1445/Tietokon eavusteinen_psykologinen_tutkimus_18 .12.2012.pdf Fisher, A., Boyle, J., Paton, J., Tomporowski, P., Watson, C., McColl, J., & Reilly, J. (2011). Effects of a physical education intervention on cognitive function in young children: Randomized controlled pilot study. BMC Pediatrics, 11(97). doi:10.1186/1471-2431-11-97 Fried, R., Hirshfeld-Becker, D., Petty, C., Batchelder, H., & Biederman, J. (2012). How informative is the CANTAB to assess executive functioning in children with ADHD? A controlled study. Journal of Attention Disorders. doi: 10.1177/1087054712457038 Gau, S. S., & Shang, C. (2010). Executive functions as endophenotypes in ADHD: Evidence from the cambridge neuropsychological test battery (CANTAB). Journal of Child Psychology and Psychiatry, 51(7), 838– 849. doi: 10.1111/j.14697610.2010.02215.x INTERNAL CONSISTENCY AND STABILITY OF THE CANTAB Gnys, J. A., & Willis, W. G. (1991). Validation of executive function tasks with young children. Developmental Neuropsychology, 7(4), 487–501. doi: 10.1080/87565649109540507 Halperin, J. M., Sharma, V., Greenblatt, E., & Schwartz, S. T. (1991). Assessment of the continuous performance test: Reliability and validity in a nonreferred sample. Psychological Assessment: A Journal of Consulting and Clinical Psychology, 3(4), 603–608. doi: 10.1037/1040-3590.3.4.603 Henry, L. A., & Bettenay, C. (2010). The assessment of executive functioning in children. Child and Adolescent Mental Health, 15(2), 110–119. doi: 10.1111/j.1475-3588.2010.00557.x Herzog, W., & Boomsma, A. (2009). Smallsample robust estimators of noncentrality-based and incremental model fit. Structural Equation Modeling, 16(1), 1-27. doi:10.1080/10705510802561279 Hu, L., & Bentler, P.M. (1999). Cutoff criteria for fit indexes in covariance structure analysis; Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6(1), 1–55. doi:10.1080/10705519909540118 Humes, G. E., Welsh, M. C., Retzlaff, P., & Cookson, N. (1997). Towers of Hanoi and London: Reliability of two executive function tasks. Assessment, 4(3), 249–257. Krikorian, R., Bartok, J., & Gay, N. (1994). Tower of London procedure: A standard method and developmental data. Journal of Clinical and Experimental Neuropsychology, 16(6), 840–850. doi: 10.1080/01688639408402697 Levaux, M. N., Potvin, S., Sepehry, A. A., Sablier, J., Mendrek, A., & Stip, E. (2007). Computerized assessment of 16 cognition in schizophrenia: promises and pitfalls of CANTAB. European Psychiatry, 22(2), 104–115. Retrieved from http://dx.doi.org/10.1016/j.eurpsy.2006. 11.004 Little, T. D., Cunningham, W. A., Shahar, G., & Widaman, K. F. (2002). To parcel or not to parcel: Exploring the question, weighing the merits. Structural Equation Modeling, 9(2), 151–173. doi :10.1207/S15328007SEM0902_1 Lowe, C., & Rabbitt, P. (1998). Test/re-test reliability of the CANTAB and ISPOCD neuropsychological batteries: Theoretical and practical issues. Neuropsychologia, 36(9), 915–923. doi: 10.1016/S0028-3932(98)00036-0 Luciana, M. (2003). Practitioner review: Computerized assessment of neuropsychological function in children: Clinical and research applications of the cambridge neuropsychological testing automated battery (CANTAB). Journal of Child Psychology and Psychiatry, 44(5), 649663. doi: 10.1111/1469-7610.00152 Luciana, M., Collins, P. F., Olson, E. A., & Schissel, A. M. (2009). Tower of London performance in healthy adolescents: The development of planning skills and associations with self-reported inattention and impulsivity. Developmental Neuropsychology, 34(4), 461–475. doi:10.1080/87565640902964540 Luciana, M., Lindeke, L., Georgieff, M., Mills, M., & Nelson, C. A. (1999). Neurobehavioral evidence for working‐ memory deficits in school‐aged children with histories of prematurity. Developmental Medicine & Child Neurology, 41(8), 521–533. doi: 10.1111/j.1469-8749.1999.tb00652.x Luciana, M., & Nelson, C. A. (1998). The functional emergence of prefrontally- INTERNAL CONSISTENCY AND STABILITY OF THE CANTAB guided working memory systems in four-to eight-year-old children. Neuropsychologia, 36(3), 273–293. doi: 10.1016/S0028-3932(97)00109-7 Luciana, M., & Nelson, C. A. (2002). Assessment of neuropsychological function through use of the cambridge neuropsychological testing automated battery: Performance in 4-to 12-yearold children. Developmental Neuropsychology, 22(3), 595–624. doi :10.1207/S15326942DN2203_3 Mammarella, I. C., Pazzaglia, F., & Cornoldi, C. (2008). Evidence for different components in children's visuospatial working memory. British Journal of Developmental Psychology, 26(3), 337–355. doi: 10.1348/026151007X236061 Milner, B. (1971). Interhemispheric differences in the localization of psychological processes in man. British Medical Bulletin; British Medical Bulletin; British Medical Bulletin, 27(3), 272–277. Miyake, A., Friedman, N.P., Rettinger, D. A., Shah, P, & Hegarty, M. (2001). How are visuospatial working memory, executive functioning, and spatial abilities related? A latent-variable analysis. Journal of Experimental Psychology: General, 130(4), 621–540. doi: 10.1037/0096-3445.130.4.621 Nunnally, J.C. & Bernstein, I.H. (1994). Psychometric Theory. (3rd ed.). New York: McGraw-Hill. Owen, A. M., Downes, J. J., Sahakian, B. J., Polkey, C. E., & Robbins, T. W. (1990). Planning and spatial working memory following frontal lobe lesions in man. Neuropsychologia, 28(10), 1021–1034. Retrieved from http://www.cambridgebrainsciences.co m/assets/default/files/pdf/Owen1990.pd f 17 Parsey, C. M., & Schmitter-Edgecombe, M. (2013). Applications of Technology in Neuropsychological Assessment. The Clinical neuropsychologist, (ahead-ofprint), 1–34. 10.1080/13854046.2013.834971 Rhodes, S. M., Riby, D. M., Matthews, K., & Coghill, D. R. (2011). Attentiondeficit/hyperactivity disorder and Williams syndrome: Shared behavioral and neuropsychological profiles. Journal of Clinical and Experimental Neuropsychology, 33(1), 147–156. doi: 10.1080/13803395.2010.495057 Robbins, T. W., James, M., Owen, A. M., Sahakian, B. J., Lawrence, A. D., McInnes, L., & Rabbitt, P. M. (1998). A study of performance on tests from the CANTAB battery sensitive to frontal lobe dysfunction in a large sample of normal volunteers: Implications for theories of executive functioning and cognitive aging. Journal of the International Neuropsychological Society, 4(5), 474– 490. Retrieved from http://journals.cambridge.org/article_S1 355617798455073 Sahakian, B., & Owen, A. (1992). Computerized assessment in neuropsychiatry using CANTAB: Discussion paper. Journal of the Royal Society of Medicine, 85(7), 399–402. Retrieved from http://www.ncbi.nlm.nih.gov/pmc/articl es/PMC1293547/pdf/jrsocmed001090031.pdf Shallice, T., & Shallice, T. (1982). Specific impairments of planning. Philosophical Transactions of the Royal Society of London.B, Biological Sciences, 298(1089), 199–209. doi: 10.1098/rstb.1982.0082 Smith, P. J., Need, A. C., Cirulli, E. T., Chiba-Falek, O., & Attix, D. K. (2013). A comparison of the Cambridge INTERNAL CONSISTENCY AND STABILITY OF THE CANTAB automated neuropsychological test battery (CANTAB) with “traditional” neuropsychological testing instruments. Journal of Clinical and Experimental Neuropsychology, (ahead-of-print), 1– 10. doi:10.1080/13803395.2013.771618 Westland, C.J. (2010). Lower bounds on sample size in structural equation 18 modeling. Electronic Commerce Research and Applications, 9(6), 476487. doi:10.1016/j.elerap.2010.07.003 Table 1 CANTAB tests measuring different dimensions of cognitive functions. Dimension of cognitive function Visual memory Executive function Attention Test Abbreviation Description Score 1. Pattern Recognition Memory PMR 2. Spatial Recognition Memory SRM No. of correct responses (max 24) No. of correct responses (max 20) 3. Spatial Span SSP 4. Stockings of Cambridge SOC 5. Intra-Extra Dimensional Set Shift IED 6. Reaction Time RTI 7. Rapid Visual Information Processing RVP Children had to remember the presented geometric patterns and discriminate them from the novel patterns. Children had to remember the location of white squares and discriminate these locations from the novel locations. Specified number of white boxes changed their color one by one and children had to reproduce the same sequence by touching the boxes in the same order the boxes changed their color. Children had to move the colored balls in the lower part of the screen to the same position in the stockings as they were in the upper part of the screen. They had specified number of moves to use. The children had to choose one of the two different dimensions: one is correct, and the other is incorrect. According to immediate feedback, they had to choose the correct pattern and learn the rule. Children had to hold down the press pad button until they saw the yellow spot flashing on one of the five circles, and then touch the middle of the circle as quickly as possible. Children had to touch the press pad button every time they discriminate the target sequence (digits 3, 5, and 7) from the digits appearing in a pseudo-random order at the rate of 100 digits per minute. Span length (max 9) No. of problems solved in minimum moves (max 12) No. of stages completed 5-choice reaction time (ms), 5-choice movement time (ms) A’ (range: .00 to 1.00; bad to good) = how well child detect the target sequences Table 2 Mean values, standard deviations (SD), and gender differences for CANTAB tests in the first assessment in 2011 Measurements in 2011 PRM no. of correct responses (max 24) SRM no. of correct responses (max 20) SSP span length (max 9) SOC no. of problems solved in minimum moves (max 12) RTI five-choice movement time (ms) RTI five-choice reaction time (ms) RVP A’ (range 0–1) IED % of children who completed at least stage 8 Boys (n=99) Mean SD 20.88 Girls (n=131) % Mean SD 2.80 20.84 16.49 1.61 6.53 Mean SD 2.29 20.86 2.52 .478 16.81 1.67 16.67 1.65 .152 1.27 6.68 1.37 6.61 1.33 .385 7.69 1.81 7.61 1.71 7.64 1.75 .746 329 74 364 87 349 83 .001 300 36 318 32 310 35 <.001 0.97 0.02 0.97 0.03 0.97 0.02 .973 76 % pa All (n=230) 73 % 74 .154 Note. Abbreviations: PRM, Pattern Recognition Memory; SRM, Spatial Recognition Memory; SSP, Spatial Span; SOC, Stockings of Cambridge; RTI, Reaction Time; RVP, Rapid Visual Information Processing; IED, Intra-Extra Dimensional Set Shift. a p-value for gender differences. INTERNAL CONSISTENCY AND STABILITY OF THE CANTAB 21 Table 3 Mean values and standard deviations (SD) for all variables for children who participated in measurements in 2011 and 2012 and effects sizes (Cohen’s d) for the repeated measures Boys (n=27) Measurements Girls (n=47) All (n=74) Mean SD d Mean SD d Mean SD d 20.96 3.39 .22 20.68 2.30 .10 20.78 2.73 .14 16.63 1.45 .02 16.70 1.79 .14 16.68 1.66 .10 6.52 1.19 .32 6.28 1.35 .55 6.36 1.29 .46 SOC no. of problems solved in minimum moves (max 12) 7.52 1.76 .43 7.38 1.69 .60 7.43 1.71 .54 RTI five-choice movement time (ms) 321 53 .05 370 83 .16 352 76 .12 RTI five-choice reaction time (ms) 304 43 -.15 319 30 .05 314 36 -.02 RVP A’ (range 0–1) .97 .02 .35 .96 .03 .40 .96 .03 .38 2011 PRM no. of correct responses (max 24) SRM no. of correct responses (max20) SSP span length (max 9) IED % of children who completed at least stage 8 2012 PRM no. of correct responses (max 24) SRM no. of correct responses (max20) SSP span length (max 9) SOC no. of problems solved in minimum moves (max 12) RTI five-choice movement time (ms) RTI five-choice reaction time (ms) RVP A’ (range 0–1) IED % of children who completed at least stage 8 75.8 73.3 74.3 Boys (n=27) Girls (n=47) All (n=74) 21.56 1.65 20.96 2.80 21.18 2.45 16.67 1.52 16.94 1.51 16.84 1.51 6.93 1.07 6.98 1.00 6.96 1.00 8.26 1.72 8.40 1.39 8.35 1.51 325 65 382 75 361 76 298 34 321 36 313 37 .98 .02 .98 .03 .98 .03 88.2 75.9 80.7 Note. Abbreviations: PRM, Pattern Recognition Memory; SRM, Spatial Recognition Memory; SSP, Spatial Span; SOC, Stockings of Cambridge; RTI, Reaction Time; RVP, Rapid Visual Information Processing; IED, Intra-Extra Dimensional Set Shift; d, Cohen’s d. INTERNAL CONSISTENCY AND STABILITY OF THE CANTAB 22 Table 4 Proportion (%) of children in 2011 and 2012 with different Spatial Span (SSP) lengths (2–9) Measurement year SSP span length 3 4 5 6 7 8 9 In 2011 1.3 1.3 22.6 20.4 23.9 24.8 5.7 In 2012 0 0 11.4 17.0 39.8 27.3 4.5 INTERNAL CONSISTENCY AND STABILITY OF THE CANTAB 23 Figure 1. The estimation results of the stability model of the Pattern Recognition Memory (PRM) test for 2011 and 2012. Standardized parameter estimates and standard errors are presented. Three parcels from 24 patterns (incorrect/correct response) of the PRM test were constructed and used as the indicator variables in the confirmatory factor analyses (PRM 1, PRM 2, PRM 3). All the factor loadings were constrained to be equal to one. INTERNAL CONSISTENCY AND STABILITY OF THE CANTAB 24 Figure 2. The stability model for the Reaction Time (RTI) test for 2011 and 2012. Standardized parameter estimates and standard errors are presented. For structural equation modeling, three subscales of the reaction time (Reac 1, Reac 2, Reac 3) and movement time (Move 1, Move 2, Move 3) at both measurement points were formed from 15 patterns of RTI. All the factor loadings were constrained to be equal to one. INTERNAL CONSISTENCY AND STABILITY OF THE CANTAB 25 Figure 3. The stability model for the Rapid Visual Information Processing (RVP) test for 2011 and 2012. Standardized parameter estimates and standard errors are presented. The RVP (multiplied by 10) of the three blocks of the test (RVP 1, RVP 2, RVP 3) were used as indicator variables in the structural equation modeling. Factor loadings were constrained to be equal across the time points. III THE ASSOCIATIONS OF OBJECTIVELY MEASURED PHYSICAL ACTIVITY AND SEDENTARY TIME WITH COGNITIVE FUNCTIONS IN SCHOOL-AGED CHILDREN by Heidi J. Syväoja, Tuija H. Tammelin, Timo Ahonen, Anna Kankaanpää & Marko T. Kantomaa. 2014. PloS one 9(7): e103559. Reproduced with kind permission by Public Library Of Science The Associations of Objectively Measured Physical Activity and Sedentary Time with Cognitive Functions in School-Aged Children Heidi J. Syva¨oja1,2*, Tuija H. Tammelin1, Timo Ahonen2, Anna Kankaanpa¨a¨1, Marko T. Kantomaa1,3 1 LIKES – Research Center for Sport and Health Sciences, Jyva¨skyla¨, Finland, 2 Department of Psychology, University of Jyva¨skyla¨, Jyva¨skyla¨, Finland, 3 Department of Epidemiology and Biostatistics, MRC–HPA Centre for Environment and Health, Imperial College London, London, United Kingdom Abstract Low levels of physical activity among children have raised concerns over the effects of a physically inactive lifestyle, not only on physical health but also on cognitive prerequisites of learning. This study examined how objectively measured and selfreported physical activity and sedentary behavior are associated with cognitive functions in school-aged children. The study population consisted of 224 children from five schools in the Jyva¨skyla¨ school district in Finland (mean age 12.2 years; 56% girls), who participated in the study in the spring of 2011. Physical activity and sedentary time were measured objectively for seven consecutive days using the ActiGraph GT1M/GT3X accelerometer. Self-reported moderate to vigorous physical activity (MVPA) and screen time were evaluated with the questions used in the ‘‘WHO Health Behavior in School-aged Children’’ study. Cognitive functions including visual memory, executive functions and attention were evaluated with a computerized Cambridge Neuropsychological Test Automated Battery by using five different tests. Structural equation modeling was applied to examine how objectively measured and self-reported MVPA and sedentary behavior were associated with cognitive functions. High levels of objectively measured MVPA were associated with good performance in the reaction time test. High levels of objectively measured sedentary time were associated with good performance in the sustained attention test. Objectively measured MVPA and sedentary time were not associated with other measures of cognitive functions. High amount of self-reported computer/video game play was associated with weaker performance in working memory test, whereas high amount of computer use was associated with weaker performance in test measuring shifting and flexibility of attention. Self-reported physical activity and total screen time were not associated with any measures of cognitive functions. The results of the present study propose that physical activity may benefit attentional processes. However, excessive video game play and computer use may have unfavorable influence on cognitive functions. Citation: Syva¨oja HJ, Tammelin TH, Ahonen T, Kankaanpa¨a¨ A, Kantomaa MT (2014) The Associations of Objectively Measured Physical Activity and Sedentary Time with Cognitive Functions in School-Aged Children. PLoS ONE 9(7): e103559. doi:10.1371/journal.pone.0103559 Editor: Yoko Hoshi, Tokyo Metropolitan Institute of Medical Science, Japan Received January 15, 2014; Accepted July 2, 2014; Published July 25, 2014 Copyright: ß 2014 Syva¨oja et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: This study was funded by Finnish Ministry of Education and Culture (101/627/2010, http://www.minedu.fi/OPM/?lang = en) and the Academy of Finland (grant 273971). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. * Email: heidi.syvaoja@likes.fi improved their cognitive performance compared to control children. Regular physical activity has been especially linked to executive functions [10]. Executive functions (also called executive control/ cognitive control) are the collection of higher-order cognitive processes controlling goal-directed actions. Core executive functions are inhibitory control (including selective attention and the inhibition of inappropriate or interfering responses), working memory and mental flexibility [11]. School-based interventions has shown that increased physical activity may improve children’s inhibitory control [12], planning ability [13] and working memory performance [14,15]. In addition, in recent studies, children with high aerobic fitness have demonstrated better inhibitory control [16–18] as well as better working memory performance [19] than less-fit children. However, evidence of the favorable effects of physical activity on cognitive functions in healthy children and adolescents is still somewhat inconsistent [20–22] and based on scarce research data [10]. Significantly, physical fitness has often been used as a proxy indicator of regular physical activity, without direct measurement Introduction In past decades, our lifestyles have become increasingly inactive [1]; only one-third of children are sufficiently active according to current physical activity recommendations [2]. Low levels of physical activity have raised concerns over the effects of a physically inactive lifestyle on children’s physical health and, recently, also on children’s learning, especially on cognitive prerequisites of learning [3]. Previous studies have shown that physical activity enhances neurocognitive function and protects against neurodegenerative diseases in elderly [4–6]. During past few years, physical activity has been linked to enhanced cognition also in children. The metaanalytic study of Sibley & Etnier [7], showed the significant overall positive association between physical activity and cognition in children. In addition, in the study of Ruiz et al. [8] leisure time physical activity was associated with better cognitive performance in adolescents. Moreover, Ardoy et al. [9] reported that children participating in high intensity physical education intervention PLOS ONE | www.plosone.org 1 July 2014 | Volume 9 | Issue 7 | e103559 Physical Activity, Sedentary Time and Cognitive Functions Table 1. Summary of the CANTAB tests used to measure different dimensions of cognitive function. Dimension of cognitive function Test Abbreviation Visual memory 1. Pattern Recognition Memory PRM Executive function 2. Spatial Span SSP 3. Intra-Extra Dimensional Set Shift IED Attention 4. Reaction Time RTI 5. Rapid Visual Information Processing RVP doi:10.1371/journal.pone.0103559.t001 participate in cognitive tests according to successful objective measurement of physical activity. If a child’s measurement did not succeed because of technical problems or the child did not remember to wear the accelerometer, they were not invited to the cognitive tests. Seven children (3 boys, 4 girls) were excluded from analysis because, according to their parents’ survey, they had physical disabilities, chronic diseases or severe learning disabilities. The final sample size used in the analyses was 224. of physical activity levels. This is particularly important, because in childhood, habitual physical activity is rarely intensive and lengthy enough to enhance aerobic fitness, and therefore, the relationship between physical activity and fitness may not be meaningful [23]. Moreover, few studies have measured a broad range of cognitive functions, highlighting the need for new studies to clarify the benefits of physical activity on different dimensions of cognitive functions. Besides physical activity, sedentary behavior and excessive media use may be associated with cognitive function in children and youth. Extensive screen time has been linked to elevated risk of attention and learning difficulties [24–27] and decreased verbal memory performance [28]. However, in other recent studies, screen time had no association with visuospatial cognition [29], and even had a positive association with enhanced attentional skills [30] and higher-developed language skills [31] in children and adolescents. Diverging research results indicates that the association of sedentary behavior and cognition is more complicated than previously believed and needs clarification. To our knowledge, no previous studies have examined the associations of objectively measured overall physical activity and sedentary time on cognitive functions in children. However, as the rates of childhood physical inactivity are increasing worldwide, it is important to better understand the potential effects of lack of physical activity and excessive sedentary time on cognitive prerequisites of learning. The purpose of this study was to examine how objectively measured and self-reported physical activity are associated with cognitive functions in school-aged children. In addition, this study aimed to determine how objectively measured sedentary time and self-reported screen time are associated with children’s cognitive functions. We hypothesized that physical activity is positively, and both sedentary time and screen time are inversely, associated with cognitive functions. Cognitive functions Cognitive functions (Table 1) were assessed using the Neuropsychological Test Automated Battery (CANTAB) (a PaceBlade Slimbook P110 tablet PC with a 12-inch touch-screen monitor and Windows XP Professional operating system, CANTABeclipse version 3). The test battery was run individually in a silent location with the guidance of trained research assistants and in accordance with the standard instructions. The execution required about 45 minutes for each individual. Visual memory was assessed with a Pattern Recognition Memory (PRM) test. PRM measures recognition memory of visual patterns in a two-choice forced discrimination paradigm. In this test, children had to remember the presented geometric patterns and discriminate them from the novel patterns. The score of the task is the number of correct responses. Executive functions were assessed with Spatial Span (SSP) and Intra-Extra Dimensional Set Shift (IED) tests. SSP measures the length of the visuospatial working memory span based on the Corsi blocks task [32]. In this test, specified number of white boxes changed their color one by one and children had to reproduce the same sequence by touching the boxes in the same order the boxes changed their color. The score of the task is the maximum number of items that the child can successfully remember in the correct order. IED is based on the Wisconsin Card Sorting test [33] and measures sifting and flexibility of attention. Specifically, it measures the ability to maintain attention to different stimuli within a relevant dimension, and shift attention to a previously irrelevant dimension. There are nine stages with increasing difficulty in this task. The children were instructed to choose one of the two different dimensions: one is correct, and the other is incorrect. According to immediate feedback, they were expected to choose the correct pattern and learn the rule. The score in this task is based on the number of stages completed. The tests assessing attention were Reaction Time (RTI) and Rapid Visual Information Processing (RVP). RTI measures children’s reaction time and speed of response to a visual target. In the unpredictable five-choice condition, a yellow spot appeared randomly in one of the five circles on the screen. Children were instructed to hold down the press pad button until they saw the yellow spot and then touch the middle of the correct circle as quickly as possible. The total score of this task is the sum of reaction time (ms) and movement time (ms). RVP is similar to the Method Ethics Statement The study was approved by the Ethics Committee of the University of Jyva¨skyla¨, and followed the principles of the Declaration of Helsinki and the Finnish legislation. Participation in the study was voluntary, and all participants had the right to drop out of the study at any time without a specific reason. Only children with a fully completed consent form (Certificate of Consent signed by a parent/guardian and the child) on the day of the first measurements were included in the study. Participants During spring 2011, 475 fifth and sixth graders were invited to participate in the study. 277 children (participation rate 58%) from five schools in the Jyva¨skyla¨ school district in Central Finland participated in the study. 230 of the 277 children were selected to PLOS ONE | www.plosone.org 2 July 2014 | Volume 9 | Issue 7 | e103559 Physical Activity, Sedentary Time and Cognitive Functions Continuous Performance Task measuring sustained attention. In this test, children had to touch the press pad button every time they discriminate the target sequence (digits 3, 5, and 7) from the digits appearing in a pseudo-random order at the rate of 100 digits per minute. The score of this task is RVP A’, which measures the child’s skill at detecting target sequences (3, 5, 7) from a pseudorandom sequence of numbers (range 0.00 to 1.00; bad to good). Internal reliability, assessed with Cronbach’s alpha reliability coefficients, was 0.49 for the RVP A’, 0.65 for the PRM number of correct responses, 0.66 for the RTI five-choice reaction time and 0.87 for the RTI five-choice movement time. For hampering tests (SSP and IED), Cronbach’s alpha could not be determined. Potential confounders Objectively measured physical activity and sedentary time Statistical analysis The parent or the child’s main caregiver filled in a questionnaire including the mother and father’s education, family income, marital status, and children’s learning difficulties and need for remedial education. The highest level of parental education, which was calculated from the mother’s and father’s education, was categorized as tertiary level education (1) and basic or upper secondary education (0). Marital status of the main caregiver was categorized as married or cohabiting (1) and divorced or single/widow (0). Children’s learning difficulties and need for remedial education were categorizes as yes (1) and no or don’t know (0). The SPSS 19.0 for Windows statistical package (SPSS (2010) IBM SPSS Statistics 19 Core System User’s Guide (SPSS Inc., Chicago, IL)) and the Mplus statistical package (Version 7; Muthe`n & Muthe`n, 1998–2012) [38] were used for the statistical analyses. Pearson’s correlation coefficients were calculated to estimate preliminary associations between objectively and subjectively measured physical activity, sedentary behavior, cognitive tests and potential confounders. Structural equation modeling was applied to examine physical activity and sedentary time in association with cognitive function. Structural equation modeling was used because it enables to estimate the measurement errors away and, therefore, increases the reliability of cognitive tests. Single scores of every problem or level of the tests were applied instead of the total score. Item parcels [39] were constructed to combine the large number of these single scores for each latent factor. These parcels were then used as indicators for latent variables. Multiple logistic regression was used to examine physical activity and sedentary time in association with cognitive function, when the outcome was dichotomous. Full information maximum likelihood (FIML) estimation with robust standard errors (MLR) was used under the assumption of data missing at random. Gender, the highest level of parental education and child’s need for remedial education were chosen to represent different aspects of potential confounders and were added to the main analysis. In order to avoid multicollinearity, highly correlated objectively measured MVPA and sedentary time were added to the model using a Cholesky factoring of the predictors [40]. The Satorra– Bentler scaled x2-test, the comparative fit index (CFI), the Tucker–Lewis Index (TLI), the root mean square error of approximation (RMSEA) and the standardized root-mean-square residual (SRMR) were used to evaluate the goodness-of-fit of the models. The model fits the data well if the p-value for the x2-test is non-significant, CFI and TLI values are close to 0.95, the RMSEA value is below 0.06 and the SRMR value is below 0.08 [41]. The ActiGraph GT1M/GT3X accelerometers with vertical axel were used to measure children’s moderate to vigorous physical activity (MVPA) and sedentary time. The accelerometer was worn on the right hip with an elastic waistband during waking hours for seven consecutive days. During bathing, swimming, and other water activities, the monitor was requested to be removed because it was not water-resistant. The ActiLife accelerometer software (ActiLife version 5; http://support.theactigraph.com/dl/ ActiLife-software) was used to collect the data. Epoch length was 10 seconds and non-wearing time 30 minutes. Customized software was used for data reduction and analysis. A cut-off value of 2,296 counts per minute was used for MVPA [34], and 100 counts per minute for sedentary time. Children were included in the analysis if they had valid data for at least 500 minutes per day on two weekdays and on one weekend day. In order to compare children, who had worn the accelerometers for different amounts of time per day, objectively measured sedentary time was expressed as percentage of daily registration time. Self-reported physical activity and screen time Physical activity and screen time were assessed with a selfreported questionnaire used earlier in the WHO Health Behavior in School-aged Children (HBSC) study [35]. Self-reported MVPA was measured with the following question: ‘‘Over the past 7 days, on how many days were you physically active for a total of at least 60 minutes per day?’’ The response categories were as follows: 0 days, 1 day, 2 days, … 7 days. There was a short description about what kind of physical activity should be taken into account when answering the question: ‘‘In the next question, physical activity is defined as any activity that increases your heart rate and makes you get out of breath some of the time.’’ Examples included running, walking quickly, rollerblading, biking, dancing, skateboarding, swimming, snowboarding, cross-country skiing, soccer, basketball, and Finnish baseball. Test-retest agreement for selfreported MVPA has been very good (ICC = 0.82) [36,37]. Selfreported screen time was evaluated with the question: ‘‘About how many hours a day do you usually a) watch television (including videos), b) play computer or video games, or c) use a computer (for purposes other than playing games, for example, emailing, chatting, or surfing the Internet or doing homework) in your free time?’’ The response options were as follows: not at all, about half an hour per day, about an hour a day, about two hours per day, … about five hours per day or more. Children responded separately for both weekdays and weekends. Test-retest agreement for watching television (ICC = 0.72–0.74) and for playing computer or video games (ICC = 0.54–0.69) has been substantial, and fair to moderate (ICC = 0.33–0.50) for using the computer [37]. Total daily screen time averages were calculated by adding these three questions, including weekdays and weekends, together. PLOS ONE | www.plosone.org Results The mean age of the children was 12.2 years and 56% of the children were girls (Table 2). 71% of children’s mothers and 56% of children’s fathers had tertiary level education, and 76% of parents were married or cohabiting. 5% of children had a diagnosed learning difficulty and 13% of children needed remedial education, according to their parent’s reports. Based on objective physical activity measurements, children had, on average, 58 minutes of MVPA per day, with no statistically significant gender difference (Table 2). However, girls spent more of their waking hours sedentary than boys (Table 2). Based on self-reports, children reported at least 60 minutes of MVPA a day for 5 days per week on average, with no significant difference between boys and girls (Table 2). Boys reported more 3 July 2014 | Volume 9 | Issue 7 | e103559 PLOS ONE | www.plosone.org 623 0.97 RTI five-choice (ms) RVP A’ (scale 0.00–1.00)b 4 1.0 Computer use (other than playing) 2.8 0.8 0.9 1.0 2.0 1.8 5.9 22.4 0.02 78 1.3 97 96 95 96 95 95 89 89 97 97 97 97 1.1 0.7 1.6 3.4 5.0 67.9 56.9 66 0.97 680 6.7 20.9 0.8 0.8 0.9 1.8 1.5 5.1 16.9 0.03 98 1.4 2.2 0.6 SD 127 127 127 127 127 127 118 127 127 127 127 127 127 N 1.0 1.0 1.6 3.7 5.0 67.2 58.4 67 0.97 655 6.6 20.9 12.2 Mean All 0.8 0.9 0.9 1.9 1.6 5.5 19.5 0.02 94 1.3 2.5 0.6 SD 223 222 223 222 222 207 207 224 224 224 224 224 224 N 0.244 ,0.001 0.982 0.015 0.923 0.013 0.231 0.891 0.946 ,0.001 0.561 0.981 0.984 pa Abbreviations: SD, standard deviation; MVPA, moderate to vigorous physical activity; PRM, Pattern Recognition Memory; SSP, Spatial Span; RTI, Reaction Time; RVP, Rapid Visual Information Processing; IED, Intra-Extra Dimensional Set Shift. a P-values for the gender differences (T-test). b A’ indicates the result in RVP test. The scale is 0.00–1.00, whereas 0.00 indicates a poor result and 1.00 a good result. c MVPA measured with the ActiGraph accelerometer using a cut-off value of 2,296 counts per minute. d Sedentary time measured by the ActiGraph accelerometer using a cut-off value 100 counts per minute and expressed as percentage of daily monitoring time (%/day). doi:10.1371/journal.pone.0103559.t002 1.4 4.0 Self-reported total screen time (h/day) Computer/video games 5.0 Self-reported MVPA (d/week with $60 min MVPA) 1.6 66.1 Objectively measured sedentary time (%/day)d TV 60.3 Objectively measured MVPA (min/day)c Measurements of physical activity and sedentary behavior 67 6.6 IED no. of children who completed the test (%) 20.9 SSP span length (scale 0–9) 12.2 0.6 12.2 97 Mean SD Mean N Girls Boys PRM no. of correct Measurements of cognitive function Age (years) Table 2. Sample characteristics according to gender and overall participants. Physical Activity, Sedentary Time and Cognitive Functions July 2014 | Volume 9 | Issue 7 | e103559 Physical Activity, Sedentary Time and Cognitive Functions goodness-of-fit statistics of the model were good (x2 (10) = 9.45 p = 0.490, CFI = 1.000, TLI = 1.010, RMSEA = 0.000, SRMR = 0.028). Objectively measured sedentary time was positively associated with RVPA’, whereas objectively measured MVPA was not associated with RVPA’ after adjusting for gender, the highest level of parental education and child’s need for remedial education (Table 3). Self-reported playing of computer/video games was negatively associated with SSP span length after adjusting for gender, the highest level of parental education and child’s need for remedial education (Table 3). The model was fully saturated. Self-reported use of computer for other purposes than playing was negatively associated with IED number of children who completed the test (Table 4). total screen time than girls, especially they spent more time playing computer or videogames than girls (Table 2). In cognitive tests, boys were faster than girls in the RTI test, but no significant gender differences were observed in performance in the PRM, SSP, RVP or IED tests. For structural equation modeling, three subscales of RTI were formed from 15 individual patterns of the RTI, and each subscale divided by 10. Each subscale loaded on the hypothesized factor. The model for the associations of objectively measured MVPA, sedentary time (SED) and performance in the Reaction Time (RTI) test is presented in Figure 1. The goodness-of-fit statistics of the model were good (x2 (10) = 14.52 p = 0.763, CFI = 0.979, TLI = 0.948, RMSEA = 0.045, SRMR = 0.018). Objectively measured MVPA was negatively associated with the RTI five-choice test score (ms), whereas objectively measured sedentary time was not associated with the RTI five-choice test score after adjusting for gender, the highest level of parental education and child’s need for remedial education (Table 3). The RVP A’ (multiplied by 10) scores of the three blocks of the test were used as indicator variables in the structural equation modeling. Each block loaded on the hypothesized factor. The model for the associations of objectively measured MVPA, sedentary time (SED) and performance in the Rapid Visual Information Processing (RVP) test is presented in Figure 2. The Discussion According to the results of this study, a high level of objectively measured MVPA was associated with good performance in the reaction time test (RTI), which measures children’s reaction time and the speed of response to a visual target. In addition, a high level of objectively measured sedentary time was associated with good performance in the sustained attention test (RVP). However, objectively measured physical activity or sedentary time were not Figure 1. Objectively measured physical activity and performance in attentional reaction time test. This figure presents the estimation results of the model for the associations of objectively measured moderate to vigorous physical activity (MVPA), sedentary time (SED) and the Reaction Time (RTI) test. Standardized parameter estimates and standard errors are presented. For structural equation modeling, three subscales of the RTI (RTI 1, RTI 2, RTI 3) (divided by 10) were formed from 15 patterns of RTI. The RTI test result is in milliseconds, where faster time indicates better performance. Confounding factors, gender (female), the highest level of parental education (tertiary level) and child’s need for remedial education (yes) were taken into account. Highly correlated objectively measured MVPA and sedentary time were added to the model as latent variables. doi:10.1371/journal.pone.0103559.g001 PLOS ONE | www.plosone.org 5 July 2014 | Volume 9 | Issue 7 | e103559 Physical Activity, Sedentary Time and Cognitive Functions Table 3. The associations between children’s cognitive processes and objectively measured physical activity, sedentary time and self-reported screen time. B SE 95% CI Pg 20.130 0.062 20.253, 20.008 0.037 20.041 0.074 20.186, 0.104 0.581 Reaction Time (RTI)a Objectively measured MVPAb Objectively measured sedentary time c Rapid Visual Information Processing (RVP)d Objectively measured MVPA 20.040 0.093 20.223, 0.143 0.669 Objectively measured sedentary time 0.305 0.078 0.153, 0.457 0.000 Self-reported viewing of TVf 20.003 0.067 20.134, 0.129 0.970 Self-reported playing of computer/video gamef 20.179 0.079 20.333, 20.024 0.023 Self-reported use of computer (other than playing)f 0.094 0.068 20.040, 0.227 0.171 Spatial Span (SSP)e Abbreviations: B, estimate; SE, standard error; CI, confidence interval; p, P-value; MVPA, moderate to vigorous physical activity. a RTI measures children’s reaction time and speed of response to a visual target in milliseconds, where faster time indicates better performance. b MVPA measured with the ActiGraph accelerometer using a cut-off value of 2,296 counts per minute and expressed as min/day. c Sedentary time measured by the ActiGraph accelerometer using a cut-off value of 100 counts per minute and expressed as percentage of daily monitoring time (%/ day). d RVP measures the sustained attention. The score of this task is RVP A’, where range is 0.00 to 1.00; bad to good. e SSP measures length of the visuospatial working memory span. The score of this task is the maximum number of items that the child can successfully reproduce. f Self-reported viewing of television, playing of computer/video games and use of computer (other than playing) are expresses as h/day. g P-values for parameter estimates. Models have been adjusted with gender (female), the highest level of parental education (tertiary level) and child’s need for remedial education. doi:10.1371/journal.pone.0103559.t003 levels of brain-derived neurotrophic factor (BDNF) [47,48] and enhance cerebrovascular function [49]; all of which may mediate the effects of physical activity on cognition. In addition, motor function has shown to be closely connected to children’s cognitive and academic skills and development [50,51], and may be an important factor driving the effects of physical inactivity on cognitive prerequisites of learning [52]. Furthermore, obesity has been related to poorer academic [52] and cognitive performance [53] and may, thereby, be one factor mediating the association of physical activity and cognition. Moreover, participation in physical activities is often a social phenomenon offering opportunities for interaction with other children and adults, and this interaction may also have a significant impact on children’s cognitive development and learning. However, this has rarely been taken into account in research [54]. Finally, physical activity may facilitate cognitive function through the cognitive demands inherent in the structure of goal-directed and engaging exercise [55]. In the present study, neither objectively measured nor selfreported physical activity was associated with visual memory, working memory, sifting and flexibility of attention or sustained attention performance. These results are in line with studies showing that physical activity is not necessary associated with all domains of cognitive functions [13–15,20,21]. These diverging results indicate the importance of future studies to define these associations. Some of the cognitive tests used in the present study were not able to differentiate healthy 12-year-old children. Neuropsychological test batteries have originally been developed to detect neurocognitive deficits, and due to that, the tests may be too easy for healthy children. In the present study, particularly in the tests of visual memory (PRM), sifting and flexibility of attention (IED) and sustained attention (RVP), children, on average, achieved very high results, which may partly explain the lack of association between physical activity and cognitive test results. Additionally, Stroth et al. [22] speculated that one reason behind associated with any other assessments of cognitive functions. High amount of self-reported playing of computer or video games was associated with weaker performance in Spatial Span (SSP) test measuring visuospatial working memory. Moreover, self-reported use of computer for other purposes than playing was negatively associated with performance in Intra-Extra Dimensional Set Shift (IED) test, which measures sifting and flexibility of attention. Selfreported physical activity, total screen time or television viewing had no association with any of the cognitive tests measuring visual memory, executive functions or attention. Physical activity and cognitive functions In our study, a high level of objectively measured physical activity was associated with better performance in attentional reaction time test. Recent study of Spitzer and Hollmann [42] supports our finding by reporting that implementation of physical activity had positive effects on children’s attention. In addition, recent intervention study of Chaddock-Heyman et al. [12] showed that physical activity enhances children’s performance in inhibitory control task requiring selective attention and inhibition of interfering responses (e.g. the Eriksen flanker task; [43]). Similarly, according to Castelli et al. [44] engagement in vigorous physical activity was positively associated with performance in inhibitory control task. However, in the studies of Fisher et al. [14], Davis et al. [13] and Puder et al. [20] physical activity intervention had no effect on children’s attentional processes. Previous studies have also reported that physically fit children outperform their less fit peers inhibitory control task [17,18,45], but also no differences in inhibitory control performance between physically fit and less fit children [22,44]. Physical activity may enhance attentional processes and other cognitive functions through different mechanisms. It has been suggested that physical activity may improve brain volume in regions supporting executive functions [19,46]; produce specific changes in the activity patterns in the brains [12,13]; increase the PLOS ONE | www.plosone.org 6 July 2014 | Volume 9 | Issue 7 | e103559 Physical Activity, Sedentary Time and Cognitive Functions Figure 2. Objectively measured sedentary time and performance in sustained attention test. This figure presents the estimation results of the model for the associations of objectively measured MVPA, sedentary time (SED) and the Rapid Visual Information Processing (RVP) test. Standardized parameter estimates and standard errors are presented. The RVP (multiplied by 10) of the three blocks of the test (RVP 1, RVP 2, RVP 3) were used as indicator variables in the structural equation modeling. The scale for the RVP test result (A’) is 0.00–1.00, where 0.00 indicates a poor result and 1.00 a good result. Confounding factors, gender (female), the highest level of parental education (tertiary level) and child’s need for remedial education (yes) were taken into account. Highly correlated objectively measured MVPA and sedentary time were added to the model as latent variables. doi:10.1371/journal.pone.0103559.g002 the result of aerobic fitness not being associated with children’s performance in cognitive control tasks may be the ceiling effect. In previous studies, aerobic fitness has often been used as a proxy measure of regular physical activity, which might contribute to diverging results: physical activity is behavior, which increases energy expenditure and occurs within a cultural context, while physical fitness is an adaptive state of the human body, affected by heritable and environmental factors and physical activity. Espe- cially in childhood, both the levels of physical activity and physical fitness may vary independently of each other due to growth, maturation and aging [23,56]. In addition, aerobic fitness measures are often confounded by adiposity and obesity in childhood [57], which are also potential factors attenuating cognitive and academic performance [53]. Moreover, the definitions, patterns and measurements of cognitive functions and physical activity have varied across different studies. Therefore, it Table 4. The associations between children’s working memory capacity and self-reported screen time. OR 95% CI Self-reported viewing of TVb 0.868 0.623, 1.210 Self-reported playing of computer/video gameb 1.321 0.943, 1.850 0.639 0.421, 0.972 Intra-Extra Dimensional Set Shift (IED) a b Self-reported use of computer (other than playing) Abbreviations: OR, odds ratio; CI, confidence interval. a IED measures sifting and flexibility of attention and was categorized as 1 children who completed the test and 0 children who did not completed the test. b Self-reported viewing of television, playing of computer/video games and use of computer (other than playing) are expresses as h/day. Model has been adjusted with gender (female), the highest level of parental education (tertiary level) and child’s need for remedial education. doi:10.1371/journal.pone.0103559.t004 PLOS ONE | www.plosone.org 7 July 2014 | Volume 9 | Issue 7 | e103559 Physical Activity, Sedentary Time and Cognitive Functions Most of the previous studies have used self-reported methods assessing sedentary behavior. In the present study, objectively measured sedentary time was positively associated with sustained attention; children who spent more time being sedentary achieved higher scores in sustained attention test. Objectively measured sedentary time is a summary measure of all kinds of sedentary behaviors, including various activities such as screen time, reading, doing homework, interaction with friends, et cetera. During some sedentary activities, such as reading and doing homework, sustained attention and distraction exclusion are needed, which may explain objectively measured sedentary time being associated with good performance in sustained attention test. In the present study, however, objectively measured sedentary time was not associated with visual memory, working memory, setsifting/mental flexibility or attentional reaction time. In addition, total screen time or television viewing were not associated with cognition. Neither did Ferguson et al. [29] observe any association between video game playing and visuospatial cognition. Moreover, in two recent studies [31,68], television viewing was not associated with cognitive and language skills after adjusting for the parent’s role in monitoring and involvement in the child’s media use. is difficult to directly compare and summarize the results of earlier studies and determine the specific benefits regular physical activity may have on cognitive functions. Our finding that only objectively measured physical activity was associated with attentional reaction time may reflect the difference between objective and subjective measurements of physical activity. Accelerometer-measured MVPA mainly illustrates cardiovascular activity with increased heart rate and respiratory frequency, while self-reported physical activity may represent different constructs and contexts. Self-reported physical activity may include skill-specific types of physical activities, which require balance and agility but hardly accumulate activity counts [58]. Our results may indicate that moderate to vigorous intensity exercise that increases cardiovascular function has benefits on attentional processes. Sedentary behavior and cognitive functions Previous studies have largely suggested that screen-based sedentary behaviors have an unfavorable effect on children’s cognition, especially on attention and learning difficulties [24–27]. The results of the present study supports the previous results by showing a negative association between self-reported computer/ video game playing and visuospatial working memory as well as negative association between self-reported use computer and shifting and flexibility of attention. Drowak et al. [28], reported also declines in verbal memory performance after computer game exposure. However, some previous studies have reported that screen-based sedentary behavior, especially videogames, is linked to enhanced cognitive skills [59–61]. According to Dye and Matthew [30], children who used to play video games had faster reaction times in attention control tests (ANT) without a notable loss in accuracy compared to non-players, indicating that action game players made faster correct responses to targets and had more resources to process distractions. Bittman et al. [31] reported that computer use was associated with higher-developed language skills. In the present study, some children spent excessive amounts of time in front of the screens on their free time: one fifth of the children reported having screen time about 5 or more hours per day. Excessive amounts of screen time may displace activities involving learning opportunities and increase children’s impulsive behavior, and eventually decrease academic skills [62]. Disadvantages, but also benefits of screen-based sedentary behavior on cognitive functions may be explained by the content of the screen time: not all screen time has an equal role in benefitting or impairing children’s cognitive skills and learning [63,64]. For example, Ennemoser & Schneider [65] reported that educational program viewing was positively, but entertainment program viewing negatively, correlated with reading speed and comprehension in children. In the study of Feng et al. [66], action game training in young adults improved performance and attenuated the gender differences favoring males in spatial attention test, while control subjects who played a non-action game showed no improvements. According to Kuhn et al. [67], video game playing may induce structural brain plasticity in the areas important to spatial navigation, strategic planning, working memory and motor performance. On the other hand, in the study of Drowak et al. [28], interactive computer game play resulted significant declines in verbal memory performance and slow wave sleep, which is important for memory consolidation, whereas viewing exciting films had no effects. Finally, it should be kept in mind that computer-based cognitive assessments may require similar cognitive skills as video and computer games, which would favor children who play a lot of video and computer games. PLOS ONE | www.plosone.org Strength and limitations of the study To our knowledge, this was the first study examining the association of objectively measured overall physical activity and sedentary time with cognitive functions in children. Conclusions regarding causality of the observed associations cannot be drawn due to the cross-sectional design. In addition, the content of sedentary time and screen-based sedentary behavior was not assessed, which limits the interpretation of the results concerning sedentary behavior. Moreover, some cognitive tests used in the present study may have been too easy for 12-year-old healthy children, and did not optimally discriminate children’s performance. Furthermore, pubertal timing, motor skills and fitness were not assessed, which limits the interpretation of the results. Future direction Studies with longitudinal designs or randomized controlled trials are needed to clarify the effects of total physical activity and sedentary behavior on cognitive prerequisites of learning. Future studies should assess what kind of physical activity or sedentary behavior affects specific kinds of cognition, as well as the mechanisms behind these associations. In the future, social interaction and context-related factors should be considered to be taken into account. In addition, the cognitive tests should be chosen so as to measure a wide range of cognitive performance. Conclusion In this study, objectively measured physical activity and sedentary time were positively associated with attentional processes, but not with other domains of cognitive functions. Self-reported computer/video game play was negatively associated with visuospatial working memory, whereas computer use was negatively associated with shifting and flexibility of attention. Selfreported physical activity and total screen time were not associated with any of the cognitive tests measuring visual memory, executive functions or attention in children. The results of the present study propose that physical activity may benefit attentional processes. However, excessive video game play and computer use may have unfavorable influence on cognitive functions. 8 July 2014 | Volume 9 | Issue 7 | e103559 Physical Activity, Sedentary Time and Cognitive Functions MTK. Wrote the paper: HJS. Gave critical input on all versions of the manuscript: HJS THT TA AK MTK. Approved the final version of the manuscript: HJS THT TA AK MTK. Author Contributions Conceived and designed the experiments: HJS THT TA MTK. Performed the experiments: HJS THT MTK. Analyzed the data: HJS THT AK MTK. Contributed reagents/materials/analysis tools: HJS THT TA AK References 23. Armstrong N, Tomkinson G, Ekelund U (2011) Aerobic fitness and its relationship to sport, exercise training and habitual physical activity during youth. Br J Sports Med 45: 849–858. doi: 10.1136/bjsports-2011-090200. 24. Johnson JG, Cohen P, Kasen S, Brook JS (2007) Extensive television viewing and the development of attention and learning difficulties during adolescence. Arch Pediatr Adolesc Med 161: 480–486. 25. Landhuis CE, Poulton R, Welch D, Hancox RJ (2007) Does childhood television viewing lead to attention problems in adolescence? Results from a prospective longitudinal study. Pediatrics 120: 532–537. doi: 10.1542/peds.2007-0978. 26. Swing EL, Gentile DA, Anderson CA, Walsh DA (2010) Television and video game exposure and the development of attention problems. Pediatrics 126: 214– 221. doi: 10.1542/peds.2009-1508. 27. Weis R, Cerankosky BC (2010) Effects of video-game ownership on young boys’ academic and behavioral functioning: A randomized, controlled study. Psychol Sci 21: 463–470. doi: 10.1177/0956797610362670. 28. Dworak M, Schierl T, Bruns T, Stru¨der HK (2007) Impact of singular excessive computer game and television exposure on sleep patterns and memory performance of school-aged children. Pediatrics 120: 978–985. doi: 10.1542/ peds.2007-0476. 29. Ferguson CJ, Garza A, Jerabeck J, Ramos R, Galindo M (2012) Not worth the fuss after all? Cross-sectional and prospective data on violent video game influences on aggression, visuospatial cognition and mathematics ability in a sample of youth. J Youth Adolesc: 42: 109–122. doi: 10.1007/s10964-012-98036. 30. Dye MW, Green CS, Bavelier D (2009) Increasing speed of processing with action video games. Curr Dir Psychol Sci 18: 321–326. doi: 10.1016/ j.neuropsychologia.2009.02.002. 31. Bittman M, Rutherford L, Brown J, Unsworth L (2011) Digital natives? New and old media and children’s outcomes. AJE 55: 161–175. 32. Milner B (1971) Interhemispheric differences in the localization of psychological processes in man. Br Med Bull 27: 272–277. 33. Mishkin M, Appenzeller T (1987) The anatomy of memory. Sci Am 256: 80–89. doi: 10.1038/scientificamerican0687-80. 34. Evenson KR, Catellier DJ, Gill K, Ondrak KS, McMurray RG (2006) Calibration of two objective measures of physical activity for children. J Sports Sci 24: 1557–1565. doi: 10.1080/02640410802334196. 35. Currie C, Zanotti C, Morgan A, Currie D, De Looze M, et al. (2012) Social determinants of health and well-being among young people. Health behaviour in school-aged children (HBSC) study. International report from the 2009/2010 survey. Health policy for children and adolescents. Available: http://www.euro. who.int/__data/assets/pdf_file/0003/163857/Social-determinants-of-healthand-well-being-among-young-people.pdf. Assessed 2014 May 6. 36. Booth M, Okely A, Chey T, Bauman A (2001) The reliability and validity of the physical activity questions in the WHO health behaviour in schoolchildren (HBSC) survey: A population study. Br J Sports Med 35: 263–267. doi: 10.1136/bjsm.35.4.263. 37. Liu Y, Wang M, Tynja¨la¨ J, Lv Y, Villberg J, et al. (2010) Test-retest reliability of selected items of health behaviour in school-aged children (HBSC) survey questionnaire in Beijing, china. BMC Med Res Methodol 10: 73. doi: 10.1186/ 1471-2288-10-73. 38. Muthe´n L, Muthe´n B, editors (1998–2012) Mplus User’s guide. Seventh edition. Los Angeles: CA: Muthe´n & Muthe´n. 39. Little TD, Cunningham WA, Shahar G, Widaman KF (2002) To parcel or not to parcel: Exploring the question, weighing the merits. Struct Equ Modeling 9: 151–173. doi:10.1207/S15328007SEM0902_1. 40. de Jong PF (1999) Hierarchical regression analysis in structural equation modeling. Struct Equ Modeling 6: 198–211. doi: 10.1080/10705519909540128. 41. Hu L, Bentler PM (1999) Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Struct Equ Modeling 6: 1–55. doi: 10.1080/10705519909540118. 42. Spitzer US, Hollmann W (2013) Experimental observations of the effects of physical exercise on attention, academic and prosocial performance in school settings. Trends Neurosci Educ 2: 1–6. doi: 10.1016/j.tine.2013.03.002. 43. Eriksen BA, Eriksen CW (1974) Effects of noise letters upon the identification of a target letter in a nonsearch task. Percept Psychophys 16: 143–149. 44. Castelli D, Hillman C, Hirsch J, Hirsch A, Drollette E (2011) FIT kids: Time in target heart zone and cognitive performance. Prev Med 52: S55–S59. doi: 10.1016/j.ypmed.2011.01.019. 45. Hillman CH, Buck SM, Themanson JR, Pontifex MB, Castelli DM (2009) Aerobic fitness and cognitive development: Event-related brain potential and task performance indices of executive control in preadolescent children. Dev Psychol 45: 114–129. doi: 10.1037/a0014437. 46. Chaddock L, Erickson K, Prakash R, VanPatter M, Voss M, et al. (2010) Basal ganglia volume is associated with aerobic fitness in preadolescent children. Dev Neurosci 32: 249–256. doi: 10.1159/000316648. 1. Pate RR, Mitchell JA, Byun W, Dowda M (2011) Sedentary behaviour in youth. Br J Sports Med 45: 906–913. doi: 10.1136/906 bjsports-2011-090192. 2. Ekelund U, Tomkinson G, Armstrong N (2011) What proportion of youth are physically active? Measurement issues, levels and recent time trends. Br J Sports Med 45: 859–865. doi: 10.1136/bjsports-2011-090190. 3. Hillman CH, Kamijo K, Scudder M (2011) A review of chronic and acute physical activity participation on neuroelectric measures of brain health and cognition during childhood. Prev Med 52: S21–S28. doi: 10.1016/ j.ypmed.2011.01.024. 4. Erickson KI, Voss MW, Prakash RS, Basak C, Szabo A, et al. (2011) Exercise training increases size of hippocampus and improves memory. Proc Natl Acad Sci U S A 108: 3017–3022. doi: 10.1073/pnas.1015950108. 5. Lautenschlager NT, Cox KL, Flicker L, Foster JK, van Bockxmeer FM, et al. (2008) Effect of physical activity on cognitive function in older adults at risk for alzheimer disease: A randomized trial. JAMA 300: 1027–1037. doi: 10.1001/ jama.300.9.1027. 6. Kramer AF, Erickson KI (2007) Effects of physical activity on cognition, wellbeing, and brain: Human interventions. Alzheimers Dement 3: S45–S51. doi: 10.1016/j.jalz.2007.01.008. 7. Sibley BA, Etnier JL (2003) The relationship between physical activity and cognition in children: A meta-analysis. Pediatr Exerc Sci 15: 243–256. 8. Ruiz JR, Ortega FB, Castillo R, Martı´n-Matillas M, Kwak L, et al. (2010) Physical activity, fitness, weight status, and cognitive performance in adolescents. J Pediatr 157: 917–922. doi: 10.1016/j.jpeds.2010.06.026. 9. Ardoy D, Ferna´ndez-Rodrı´guez J, Jime´nez-Pavo´n D, Castillo R, Ruiz J, et al. (2014) A physical education trial improves adolescents’ cognitive performance and academic achievement: The EDUFIT study. Scand J Med Sci Sports 24: e52–e61. doi: 10.1111/sms.12093. 10. Guiney H, Machado L (2013) Benefits of regular aerobic exercise for executive functioning in healthy populations. Psychon Bull Rev 20: 73–86. doi: 10.3758/ s13423-012-0345-4. 11. Diamond A (2013) Executive functions. Annu Rev Psychol 64: 135–168. doi: 10.1146/annurev-psych-113011-143750. 12. Chaddock-Heyman L, Erickson KI, Voss MW, Knecht AM, Pontifex MB, et al. (2013) The effects of physical activity on functional MRI activation associated with cognitive control in children: A randomized controlled intervention. Front Hum Neurosci 7. doi: 10.3389/fnhum.2013.00072. 13. Davis C, Tomporowski P, McDowell J, Austin B, Miller P, et al. (2011) Exercise improves executive function and achievement and alters brain activation in overweight children: A randomized, controlled trial. Health Psychol 30: 91–98. doi: 10.1037/a0021766. 14. Fisher A, Boyle J, Paton J, Tomporowski P, Watson C, et al. (2011) Effects of a physical education intervention on cognitive function in young children: Randomized controlled pilot study. BMC Pediatr 11: 97. doi: 10.1186/14712431-1-97. 15. Kamijo K, Pontifex M, O’Leary K, Scudder M, Wu C, et al. (2011) The effects of an afterschool physical activity program on working memory in preadolescent children. Dev Sci: 1–13. doi: 10.1111/j.1467-7687.2011.01054.x. 16. Chaddock L, Hillman CH, Pontifex MB, Johnson CR, Raine LB, et al. (2012) Childhood aerobic fitness predicts cognitive performance one year later. J Sports Sci 30: 421–430. doi: 10.1080/02640414.2011.647706. 17. Chaddock L, Erickson KI, Prakash RS, Voss MW, VanPatter M, et al. (2012) A functional MRI investigation of the association between childhood aerobic fitness and neurocognitive control. Biol Psychol 89: 260–268. doi: 10.1016/ j.biopsycho.2011.10.017. 18. Pontifex MB, Raine LB, Johnson CR, Chaddock L, Voss MW, et al. (2011) Cardiorespiratory fitness and the flexible modulation of cognitive control in preadolescent children. J Cogn Neurosci 23: 1332–1345. doi: 10.1162/ jocn.2010.21528. 19. Chaddock L, Erickson K, Prakash R, Kim J, Voss M, et al. (2010) A neuroimaging investigation of the association between aerobic fitness, hippocampal volume, and memory performance in preadolescent children. Brain Res 1358: 172–183. doi: 10.1016/j.brainres.2010.08.049. 20. Puder J, Marques-Vidal P, Schindler C, Zahner L, Niederer I, et al. (2011) Effect of multidimensional lifestyle intervention on fitness and adiposity in predominantly migrant preschool children (ballabeina): Cluster randomised controlled trial. BMJ 343: d6195. doi: 10.1136/bmj.d6195. 21. Reed JA, Maslow AL, Long S, Hughey M (2012) Examining the impact of 45 minutes of daily physical education on cognitive ability, fitness performance, and body composition of African American youth. J Phys Act Health 10: 185–197. 22. Stroth S, Kubesch S, Dieterle K, Ruchsow M, Heim R, et al. (2009) Physical fitness, but not acute exercise modulates event-related potential indices for executive control in healthy adolescents. Brain Res 1269: 114–124. doi:10.1016/ j.brainres.2009.02.073. PLOS ONE | www.plosone.org 9 July 2014 | Volume 9 | Issue 7 | e103559 Physical Activity, Sedentary Time and Cognitive Functions 57. Rowland T (2013) Oxygen uptake and endurance fitness in children, revisited. Pediatr Exerc Sci 25: 508–514. 58. Syva¨oja HJ, Kantomaa MT, Ahonen T, Hakonen H, Kankaanpa¨a¨ A, et al. (2013) Physical activity, sedentary behavior, and academic performance in Finnish children. Med Sci Sports Exerc 45: 2098–2104. doi: 10.1249/ MSS.0b013e318296d7b8. 59. Boot WR, Blakely DP, Simons DJ (2011) Do action video games improve perception and cognition? Frontiers in Psychology 2. doi: 10.3389/ fpsyg.2011.00226. 60. Spence I, Feng J (2010) Video games and spatial cognition. Rev Gen Psychol 14: 92–104. doi: 10.1037/a0019491. 61. Granic I, Lobel A, Engels RC (2013) The benefits of playing video games. Am Psychol 69: 66–78. doi: 10.1037/a0034857. 62. Shin N (2004) Exploring pathways from television viewing to academic achievement in school age children. J Genet Psychol 165: 367–381. doi: 10.3200/GNTP.165.4.367-382. 63. Kirkorian HL, Wartella EA, Anderson DR (2008) Media and young children’s learning. Future Child 18: 39–61. doi: 10.1353/foc.0.0002. 64. Schmidt ME, Vandewater EA (2008) Media and attention, cognition, and school achievement. Future Child18: 63–85. doi: 10.1353/foc.0.0004. 65. Ennemoser M, Schneider W (2007) Relations of television viewing and reading: Findings from a 4-year longitudinal study. J Educ Psychol 99: 349–368. doi: 10.1037/0022-0663.99.2.349. 66. Feng J, Spence I, Pratt J (2007) Playing an action video game reduces gender differences in spatial cognition. Psychol Sci 18: 850–855. doi: 10.1111/j.14679280.2007.01990.x. 67. Ku¨hn S, Gleich T, Lorenz R, Lindenberger U, Gallinat J (2013) Playing super mario induces structural brain plasticity: Gray matter changes resulting from training with a commercial video game. Mol Psychiatry 19: 265–271. doi: 10.1038/mp.2013.120. 68. Munasib A, Bhattacharya S (2010) Is the ‘Idiot’s Box’raising idiocy? Early and middle childhood television watching and child cognitive outcome. Econ Educ Rev 29: 873–883. doi: 10.1016/j.econedurev.2010.03.005. 47. Gomez-Pinilla F, Hillman C (2013) The influence of exercise on cognitive abilities. Compr Physiol 3: 403–428. doi: 10.1002/cphy.c110063. 48. Hopkins M, Davis F, Vantieghem M, Whalen P, Bucci D (2012) Differential effects of acute and regular physical exercise on cognition and affect. Neuroscience 215: 59–68. doi: 10.1016/j.neuroscience.2012.04.056. 49. Brown AD, McMorris CA, Longman RS, Leigh R, Hill MD, et al. (2010) Effects of cardiorespiratory fitness and cerebral blood flow on cognitive outcomes in older women. Neurobiol Aging 31: 2047–2057. doi:10.1016/j.neurobiolaging.2008.11.002. 50. Haapala EA, Poikkeus A, Tompuri T, Kukkonen-Harjula K, Leppa¨nen P, et al. (2013) Associations of motor and cardiovascular performance with academic skills in children. Med Sci Sports Exerc: In press. doi: 10.1249/ MSS.0000000000000186. 51. Iverson JM (2010) Developing language in a developing body: The relationship between motor development and language development*. J Child Lang 37: 229. doi: 10.1017/S0305000909990432. 52. Kantomaa MT, Stamatakis E, Kankaanpa¨a¨ A, Kaakinen M, Rodriguez A, et al. (2013) Physical activity and obesity mediate the association between childhood motor function and adolescents’ academic achievement. PNAS 110: 1917–1922. doi:10.1073/pnas.1214574110/2/DCSupplemental. 53. Burkhalter TM, Hillman CH (2011) A narrative review of physical activity, nutrition, and obesity to cognition and scholastic performance across the human lifespan. Adv Nutr 2: 201S–206S. doi: 10.3945/an.111.000331. 54. Hillman CH, Erickson KI, Kramer AF (2008) Be smart, exercise your heart: Exercise effects on brain and cognition. Nat Rev Neurosci 9: 58–65. doi:10.1038/nrn2298. 55. Best JR (2010) Effects of physical activity on children’s executive function: Contributions of experimental research on aerobic exercise. Dev Rev 30: 331– 351. doi: 10.1016/j.dr.2010.08.001. 56. Malina RM (1996) Tracking of physical activity and physical fitness across the lifespan. Res Q Exerc Sport 67: S48–S57. doi 10.1080/ 02701367.1996.10608853. PLOS ONE | www.plosone.org 10 July 2014 | Volume 9 | Issue 7 | e103559
© Copyright 2024