key: cord-0025473-tft81jud authors: Markevych, Iana; Orlov, Natasza; Grellier, James; Kaczmarek-Majer, Katarzyna; Lipowska, Małgorzata; Sitnik-Warchulska, Katarzyna; Mysak, Yarema; Baumbach, Clemens; Wierzba-Łukaszyk, Maja; Soomro, Munawar Hussain; Compa, Mikołaj; Izydorczyk, Bernadetta; Skotak, Krzysztof; Degórska, Anna; Bratkowski, Jakub; Kossowski, Bartosz; Domagalik, Aleksandra; Szwed, Marcin title: NeuroSmog: Determining the Impact of Air Pollution on the Developing Brain: Project Protocol date: 2021-12-28 journal: Int J Environ Res Public Health DOI: 10.3390/ijerph19010310 sha: 0e0427cff9c0f738e17a11619bf0b80a71c8de95 doc_id: 25473 cord_uid: tft81jud Exposure to airborne particulate matter (PM) may affect neurodevelopmental outcomes in children. The mechanisms underlying these relationships are not currently known. We aim to assess whether PM affects the developing brains of schoolchildren in Poland, a country characterized by high levels of PM pollution. Children aged from 10 to 13 years (n = 800) are recruited to participate in this case–control study. Cases (children with attention deficit hyperactivity disorder (ADHD)) are being recruited by field psychologists. Population-based controls are being sampled from schools. The study area comprises 18 towns in southern Poland characterized by wide-ranging levels of PM. Comprehensive psychological assessments are conducted to assess cognitive and social functioning. Participants undergo structural, diffusion-weighted, task, and resting-state magnetic resonance imaging (MRI). PM concentrations are estimated using land use regression models, incorporating information from air monitoring networks, dispersion models, and characteristics of roads and other land cover types. The estimated concentrations will be assigned to the prenatal and postnatal residential and preschool/school addresses of the study participants. We will assess whether long-term exposure to PM affects brain function, structure, and connectivity in healthy children and in those diagnosed with ADHD. This study will provide novel, in-depth understanding of the neurodevelopmental effects of PM pollution. Air pollution has received increased global public attention in recent years as presenting a major risk to human health [1] . According to the estimates of the World Health on these systems will be investigated using behavioural tests, including IQ. To meet the study aim, we assembled a multidisciplinary team of experts comprising air pollution modellers, environmental epidemiologists, neuroscientists, and clinical psychologists. We selected towns in southern Poland, based on their levels of particulate air pollution and population size, that are located within two hours' drive to the magnetic resonance imaging (MRI) scanning centre in Kraków. The particulate air pollution level of each candidate town (n = 202) was classified as high, medium, or low based on the populationweighted median air pollution level inferred from 1 km × 1 km interpolated PM 2.5 data for the year 2015 [34] . We chose small and large towns across different levels of air pollution to minimize urbanicity-related confounding. Towns with more than 90,000 inhabitants were classified as large; others were classified as small. Accordingly, we selected 18 towns (Table 1, Figure 1 ). We are recruiting comparable numbers of children (cases and controls) for each combination of air pollution level and town population size (Table 1) . Within each such combination, the number of children recruited per town is approximately proportional to the town's population size. Recruitment of study participants began in October 2020, with the aim of having 800 children complete all study procedures by the end of 2022. We are recruiting one ADHD case (intended number n = 267) per two population controls (intended number n = 533). Cases with ADHD diagnosis according to the 11th revision of the International Classification of Diseases and Related Health Problems (ICD-11) [35] are recruited by field psychologists in the study towns. Children at risk of ADHD are referred to the psychologists either by cooperative school psychologists or teachers, or by paediatricians, psychologists, psychiatrists, or neurologists from psychological-educational counselling centres and mental health centres. In addition, cases are also recruited from families who voluntarily contacted us after seeing our recruitment advertisements. We use a two-step model to verify and eliminate ADHD diagnosis in cases and controls, respectively (see Section 2.6). Firstly, a comprehensive psychological assessment is administered and evaluated by field psychologists. Secondly, three consultant clinical psychologists review the assessments to verify the diagnoses according to the ICD-11. Cases that do not meet the ICD-11 criteria are excluded from the study; controls who meet the ADHD diagnostic criteria are recruited into the case arm. All study participants are fluent Polish speakers, aged between 10 and 13 years, have average or above average intelligence (i.e., attend non-specialised schools), are in grades IV-VI, and attend school in the selected towns. Exclusion criteria for all are diagnosed intellectual disability, neurological or comorbid psychiatric disorders, other serious medical conditions, as well as contraindications to MRI. Children born before the 35th week or after the 40th week of gestation, or with birth weight <2500 g or with an Apgar score <8 are also excluded. As we are interested in the long-term effects of air pollution, children who resided outside of Poland during the last year are also excluded, as are children whose legal guardians are not fluent in Polish. A team of field psychologists with at least five years of clinical experience was recruited to enlist potential cases and perform psychological evaluation of all children. Prior to beginning their work, all psychologists received accredited training in the use of psychological tests utilised in the study (see Section 2.6). They were also trained in the administration of two behavioural tests (see Section 2.7), as well as in data protection and privacy laws. Population controls are recruited from randomly selected larger non-specialized primary schools within each of the study towns (one to five schools per town). Lists of candidate schools were retrieved from the Boards of Education (Kuratoria Oświaty). Within each of the selected schools, classes are randomly selected in grades IV-VI. Within each class, one to five pupils are selected at random. The sampling is defined a priori using randomized lists defining which classes in every grade and which pupils in every selected class to approach and in what order. The randomization was implemented using the sample.int() function in R statistical software [36] . Headteachers of selected schools are contacted and asked to support the study's recruitment campaign and sign a formal collaborative agreement. Participating schools are promoted on the project website, and are offered workshops on neurodevelopment and on the study results. Consent for contact from the legal guardians is obtained by the schools. Families who give their consent are contacted by telephone by research assistants. The assistants explain the study aims and procedures, confirm study eligibility, and invite the legal guardians to participate in the study. If a school does not agree to participate, the next school in the randomized list of schools is contacted. If the legal guardians of a selected child do not consent to being contacted, the next family from the randomized list of pupils is contacted. Where insufficient numbers of pupils are recruited in each grade, additional pupils are selected from the next class in the randomized list of classes (for that grade) using the same procedure until the desired number of pupils are recruited. All psychological and behavioural assessments are performed during three meetings with the field psychologist. Each meeting lasts about one hour and twenty minutes to two hours, and both the child and their legal guardian are present. At the final meeting, the results of the assessments are explained to the child and to their guardian, a formal psychological evaluation is provided. The MRI scan is completed in Kraków within three months of the psychological evaluation. Typical MRI sessions last up to one and a half hours. After the MRI session, families are rewarded with vouchers equal to 270 PLN (about 60 EUR at the time of writing) and are reimbursed for travel expenses. The children receive t-shirts printed with MRI images of their brains. Figure 2 illustrates the progress in the recruitment of controls at the end of the first study year (June 2021). A total of 370 children have undergone psychological evaluation, of whom 236 have also completed the MRI scan. Figure 3 depicts the number of cases and controls tested in each town at the end of August 2021. Behavioural testing is conducted on HP laptops with 15.6" screens, set up identically for each field psychologist. The laptops are encrypted using BitLocker and synchronised with a secure OneDrive cloud folder at the Jagiellonian University that collects raw data from every test performed. To minimise the impact of the study on the environment and to reduce transcription from paper to computer, the psychological data are-wherever possible-collected digitally using online questionnaires implemented in survey tools developed by Qualtrics (https: //www.qualtrics.com, accessed on 23 December 2021). The raw data are processed and prepared for import in R and Python. To allow easy data selection and extraction of variables of interest by data analysts, a secure user website will be implemented that will interface with an SQLite database. Two questionnaires have been developed for the study: a General Questionnaire, and an Address Questionnaire. The General Questionnaire is a Qualtrics-based tool filled in by a legal guardian that collects data on confounders, effect modifiers, and additional exposures. Some items were adopted from the existing studies [37] [38] [39] [40] , and some of the items were invented for the purpose of NeuroSmog. Briefly, information on sociodemographics, pregnancy and early life, general health, habits, academic performance, home, and neighbourhood environments of the child is being collected (see Supplementary File S1 for a paper-adopted version of the questionnaire). Non-response options are always presented as a variation of "not applicable", "don't know", "don't wish to respond", and at the same time, all responses are forced. The General Questionnaire was validated on a small sample of volunteers. The Address Questionnaire collects full prenatal and postnatal residential history of the study participants, as well as preschool and school addresses, hours spent in preschools and at schools, and the transportation mode used (see Supplementary File S2). It is a pencil-and-paper questionnaire. The collected addresses are being geocoded and will be used to assign air pollution estimates, as well as other geographical exposures. Cognitive functioning is examined using the Stanford-Binet Intelligence Scales, 5th edition (SB5, [41] and tests from the Diagnostic Battery for Cognitive Functions Evaluation PU1, [42] ). These tests measure attention, working memory, and executive function, which represent domains relevant to both classroom learning and ADHD [43] . Guardians complete the Polish adaptations of the Third Edition of Conners' Rating Scales [44] and Child Behaviour Checklist (CBCL) for Ages 6-18 [45] , providing information on ADHD and other externalising and internalising behavioural issues. Children complete the Polish adaptation of Youth Self-Report (YSR) [46] , which is a child-completed version of the CBCL [45, 47] . Social functioning is assessed using two questionnaires. The first is the Polish adaptation of the Family Adaptation and Cohesion Evaluation Scales (FACES-IV) questionnaire (Skala Oceny Rodziny; SOR) [48] , which measures family function, communication, and family satisfaction. The SOR is completed by both guardians and children. The second is the Polish Siblings Relationship Questionnaire (Kwestionariusz Relacji z Rodzeństwem W Okresie Adolescencji; KRR) [49] , which is completed only by children with siblings. The attention network task (ANT) measures the efficiency of the three attentional networks-orienting, alerting, and executive functioning [50] . The ANT demonstrates high immediate test-retest reliability, and scores are not correlated across individuals, suggesting that each network efficiency can be measured somewhat independently [51] . We use the children's version of the ANT, with timing and task stimuli operationalised identically to those used in [50] with the exception that, since our subjects are older, cue duration is set to 100 ms (as in the adult version). The continuous performance test (CPT) is a variant of the Go/NoGo task which allows quantification of deficits in attentiveness, impulsivity, activation/arousal, and vigilance. The CPT has high three-months test-retested reliability [52] and has been validated in children with ADHD and in healthy controls [53] . During the task, geometric figures appear (250 ms) on the computer screen in 18 consecutive blocks of 20 trials (360 trials total). The geometric figures are identical to those used subsequently in the functional MRI (fMRI) variant of the Go/NoGo task. The interstimulus interval (ISI) is set to either 1, 2, or 4 s during each block. Participants are asked to press the spacebar in response to each letter presentation, except for the square and the circle [54] . The proportions of Go to NoGo trials are set at 80% and 20%, respectively [55] . All children complete practice runs of both tests until their performance reaches 90% accuracy. All neuroimaging data are acquired at the Małopolska Centre of Biotechnology, Jagiellonian University in Kraków, Poland, on a Siemens MAGNETOM Skyra 3T MRI scanner (Erlangen, Germany) using a 64-channel head coil. Participants are familiarised with the scanner and trained to remain still using a mock scanner (https://pstnet.com/, accessed on 23 December 2021). The mock scanning session lasts for 10 min. During the scan, head movements are constrained by inflatable Pearltec pillows (www.pearl-technology.ch, accessed on 23 December 2021). The protocol includes T1-weighted (T1-w), T2-weighted (T2-w), and magnetizationprepared 2 rapid acquisition gradient echo (MP2RAGE) structural images, diffusionweighted imaging (DWI), two resting-state fMRI (rsfMRI), and two task fMRI runs (see Table 2 for detailed acquisition parameters and order). The scanning protocol, including instructions, lasts for one hour. The acquisition sequences for T1-w, T2-w, and fMRI neuroimaging data were adopted from the Adolescent Brain Child Development (ABCD) project [56] . The DWI acquisition sequence was adapted from the UK Biobank (www.ukbiobank.ac. uk, accessed on 23 December 2021) to fit the study scanner and includes multiple b-value volume acquisitions. For the fMRI and DWI acquisitions, we use multi-band accelerated echo planar imaging (EPI) pulse sequences developed by the Center for Magnetic Resonance Research at the University of Minnesota [57, 58] . The T1-w sequence provides a better contrast-to-noise ratio in white matter and is used for cortical and subcortical segmentation of the brain [59] . The T2-w sequence allows for the discrimination of structural differences in cerebral fluid-filled regions. The combined use of T1-w/T2-w allows the generation of ratio-based cortical myelin maps [60] . It also provides the anatomical reference for rsMRI and task fMRI data, including anteriorposterior commissure alignment. For improved quality, the T1-w and T2-w protocols include volumetric navigators for prospective motion correction [61] . MP2RAGE achieves a spatially uniform contrast [62] and is used for morphometric analyses of brain anatomy and for construction of cortical myelin maps. DWI enables visualisation and characterisation of white matter tracts [63] . rsfMRI imaging allows for the inference of the intrinsic organization of large-scale brain networks [64] . Throughout the structural, MP2RAGE, and DWI acquisitions, participants watch neutral-valence movies (e.g., nature documentaries on the life of birds); throughout the rsfMRI acquisition, participants are instructed to stay awake, to keep still, and to blink normally while looking at a fixation cross. DWI-Diffusion-weighted imaging; fMRI-Functional magnetic resonance imaging; FOV-field of view; MP2RAGE-Magnetization-prepared 2 rapid acquisition gradient echo; rsfMRI-Resting-state functional magnetic resonance imaging; T1-w-T1-weighted; T2-w-T2-weighted; TE-echo time; TI-inversion time; TR-repetition time. Finally, task fMRI analyses can target task-dependent whole-brain activity and functional connectivity [65, 66] . Based on our hypotheses, we chose an fMRI task that measures response inhibition as a measure of executive functioning, as inhibitory control is positively associated with cognitive flexibility and problem-solving skills and, thus, represents an important functional domain during brain maturation [67] . In addition, diminished inhibitory control is associated with disorders of impulse control, including ADHD [68, 69] . This task is identical to the conditioned approach response inhibition task (CARIT) used in the Human Connectome Project-Development (HCP-D) project [70] and is a variant of the Go/NoGo task. The task allows mapping of differential neuronal activity when response inhibition demands are high (NoGo trials) as compared to free prepotent motor execution (Go trials) [52] . The task uses a rapid event-related fMRI design with jittered inter-trial intervals (1000-4500 mm) and randomized inter-target intervals to optimize statistical efficiency [71] . During each run, participants view shape stimuli (n = 92) and are instructed to press a button as quickly as possible in response to every shape (Go; n = 68) except for a circle and a square (NoGo; n = 24) [70, 72] . Region-of-interest measures will be exported from appropriate MRI analysis packages, including FreeSurfer (http://surfer.nmr.mgh.harvard.edu/; accessed on 23 December 2021), Statistical Parametric Mapping (SPM) 2 (www.fil.ion.ucl.ac.uk/spm, accessed on 23 December 2021), and the Oxford Centre for Functional MRI of the Brain Software Library (FSL; https://fsl.fmrib.ox.ac.uk, accessed on 23 December 2021), and analysed in R [36] using logistic regression, with common best statistical practices such as applying corrections for multiple comparisons. Whole-brain measures will be calculated within the MRI analysis packages themselves and statistical tests appropriate to each imaging modality will be applied. For example, for task fMRI data, we will apply a voxel-wise threshold of either p < 0.01 or p < 0.001 and an appropriate cluster-level threshold (p < 0.05, false discovery rate, FDR). Three health-relevant air pollutants are considered for this study: PM 2.5 , PM 10 , and NO 2 . Annual and monthly grids of air pollutant concentrations will be created for the study area for the period from 2007 to 2022 using hybrid land use regression (LUR) models, following methods developed by de Hoogh and colleagues [73, 74] . Interestingly, the research evaluating the performance of various regression models, regularization techniques, and machine learning methods for air pollution spatial modelling has shown relatively small differences in the prediction of accuracy in air concentration between various algorithms [75] . The LUR models are estimated and validated using measurements from the air quality stations for each of the air pollutants and for each year separately, and will also include monthly LUR models for the prenatal period. Routine measurements of these air pollutants are made by Polish Chief Inspectorate of Environmental Protection via a network of air quality monitoring stations [76] . The locations of the monitoring stations (n = 179) in the study area are illustrated in Figure 1 . Measured air pollution concentrations between 2007 and 2019 are depicted in Figure 4 . Predictor variables for the LUR models are summarized in Table 3 , and detailed in Supplementary Table S1. We calculate atmospheric dispersion of air pollutants according to meteorological conditions using the methodology described in [77, 78] , providing 1-h concentrations of pollutants. We use satellite images to inform about land use changes. Data processing is completed with QGIS geographic information system (GIS) software [79] . Land cover predictor variables are calculated for seven girds at the following resolutions:~4 km × 4 km, 2 km × 2 km, 1 km × 1 km, 500 m × 500 m, 250 m × 250 m, 125 m × 125 m, and 62.5 m × 62.5 m. Exploratory and correlation analyses of average air pollutant concentrations from the monitoring stations and the predictor variables are performed to assess temporal and spatial distribution patterns. Optimal resolution for each type of variable applied in the LUR air quality modelling is determined based on physical interpretability and supported with statistical analyses, including machine learning. Predictor variables are included in the LUR model only if they adhere to the predefined direction of effect. Models are validated with out-of-sample and cross-validations methods using data from the monitoring stations. Model assumptions are checked through the analysis of residuals, as well as visual inspection of quantile-quantile plots. Log transformations are tested to minimize the effect of non-normal errors, and p-values are calculated assuming normal distributions of errors. Kriging of air pollution concentrations will be performed, as it is expected to significantly improve model performance of linear regression algorithms for spatial data [75] . All LUR modelling is conducted in R statistical software [36] . Finally, estimates resulting from the LUR models will be assigned to all residential and preschool/school addresses of each study participant. Prenatal, early postnatal, concurrent, and life-long air pollution estimates will be calculated. The main outcomes resulting from psychological evaluation, behavioural testing, and neuroimaging are listed in Table 4 . We are planning to publish several papers using prenatal, early life, and lifetime PM 2.5 , PM 10 , and NO 2 as exposures. The papers will be grouped by outcome domain: Structural brain measures 2. Fractional anisotropy 3. Brain measures of inhibitory and attentional functions assessed with task fMRI 4. Resting state connectivity 5. Measures of attention assessed by behavioural tasks 6. ADHD and other externalizing behaviours 7. IQ Opinions of assessing psychologists, Conners 3, PU1, SB5, CBCL, and validation ADHD diagnosis ADHD-Attention deficit hyperactivity disorder; ANT-attention network test; BOLD-Blood oxygen level dependent effect; CARIT-Conditioned approach response inhibition task; CBCL-Child Behaviour Checklist; CPT-Continuous performance test; DWI-CDiffusion-weighted imaging; fMRI-Functional magnetic resonance imaging; IQ-Intelligence quotient; MP2RAGE-Magnetization-prepared 2 rapid acquisition gradient echo; NODDI-Neurite orientation dispersion and density; PU1-Bateria diagnozy funkcji poznawczych PU1: Pamięćuwaga-funkcje wykonawcze (Diagnostic Battery for Cognitive Functions Evaluation); SB5-Stanford-Binet Intelligence Scales, 5th edition; T1-w-T1-weighted; T2-w-T2-weighted; YSR-Youth Self-Report. Regression analyses will be used to analyse each exposure-outcome pair separately. We will individually assess associations with every pollutant at every exposure window. We will also consider adjusting our models for residuals of co-pollutants, given that PM and NO 2 tend to be strongly correlated. Confounder sets will be selected using directed acyclic graphs (DAGs) [80] to avoid overadjustment. The primary objective of the NeuroSmog study is to combine state-of-the-art multimodal neuroimaging, psychological assessment, environmental epidemiology, and air pollution modelling to determine the impact of air pollutants on neurophysiological and behavioural outcomes in healthy children and in atypically developing children diagnosed with ADHD. NeuroSmog is based on an ethnically, racially, and culturally homogeneous Polish population, which reduces potential confounding. NeuroSmog is the first large-scale study to provide collateral measures of brain structure, function, and connectome, with periodically stratified and lifespan (including prenatal) exposures to air pollutants, and comprehensive child psychological assessments. We use a single MRI scanner and a narrow age range of participants to eliminate multisite acquisition confounding and to reduce epiphenomena related to developmental and hormonal changes in children. Additionally, we have adapted our protocol to fit two other large neuroimaging studies, the ABCD and the HCP-D, and we will make our data openly available to the scientific community providing a rich and easily comparable resource. A particular novelty of NeuroSmog is the inclusion of an fMRI task that directly tests the functioning of neural systems for inhibitory control that are putatively affected by air pollution exposure. The few neuroimaging air pollution studies conducted to date were done in cities with relatively clean air. Performing this work in southern Poland allows us to investigate the impact of air pollution at concentrations and ranges greater than studied previously, thus increasing the chances of discovering associations, if present. Compared to existing similar studies, the project is unique in terms of its case-control design. Children with ADHD are hypothesized to be more vulnerable to the effects of air pollution. Finally, high-resolution air pollution grids will be produced for the study area using state-of-the-art techniques. We are, thus, confident that the data generated by our project, when combined with appropriate analysis and interpretation, will yield new insights into factors that impact or alter neurodevelopmental trajectories in children. The COVID-19 pandemic, and disruptions to work, movement, and education that are associated with it, have greatly complicated the fieldwork component of the study. Even under normal conditions, the logistics of managing fieldwork across multiple centres are challenging. In general, all fieldwork has required more effort than originally anticipated and is taking longer than planned. The pandemic has also impacted negatively on response rates: 50% of contacted schools declined to collaborate and outreach towards the remaining 50% remains challenging. Use of MRI in research presents additional challenges. Some members of the public distrust MRI and, on occasion, it has been difficult to motivate families of both cases and controls to participate in the study. A mixed attitude towards research involving underaged participants, exacerbated by the recent rise of anti-vaccine sentiment, also poses a challenge. Although we will be able to assess lifelong exposure to air pollution in our study, we are only able to investigate the associations with a limited number of air pollutants. Our PM measures are gross, and we have no knowledge of its actual composition, which could vary considerably across Poland. There are a host of environmental stressors that could potentially affect neurodevelopment and be associated with air pollutants (and therefore cause confounding). The study would be better able to capture some of the wide array of environmental chemicals which might affect neurodevelopment if we had biological samples. Finally, as with all case-control studies, we anticipate that recall bias may affect our analyses, particularly in terms of early life factors and residences. Author Contributions: I.M. designed the study, conceptualized the fieldwork, and is overseeing it, coordinating and executing the data management, consulting on the air pollution modelling, drafted the first version of the study protocol, revised the draft, and approved the final manuscript. N.O. is coordinating and executing the data management, drafted the first version of the study protocol, revised the draft, and approved the final manuscript. J.G. designed the study and obtained the necessary funds, is consulting on the air pollution modelling, revised the study protocol, and approved the final manuscript. K.K.-M. is responsible for the air pollution modelling, drafted the first version of the study protocol, revised the draft, and approved the final manuscript. M.L. designed the study, conceptualized the fieldwork and is overseeing it, revised the study protocol, and approved the final manuscript. K.S.-W. designed the study, conceptualized the fieldwork and is overseeing it, revised the study protocol, and approved the final manuscript. Y.M. conceptualized the fieldwork and is overseeing it, is coordinating and executing the data management, drafted the first version of the study protocol, revised the draft, and approved the final manuscript. C.B. designed the study, conceptualized the fieldwork and is overseeing it, is coordinating and executing the data management, drafted the first version of the study protocol, revised the draft, and approved the final manuscript. M.W.-Ł. conceptualized the fieldwork and was overseeing it, was coordinating and executing the data management, revised the study protocol, and approved the final manuscript. M.H.S. drafted the first version of the study protocol, revised the draft, and approved the final manuscript. M.C. conceptualized the fieldwork and is overseeing it, is executing the data management, revised the study protocol, and approved the final manuscript. B.I. designed the study, conceptualized the fieldwork and oversees it, revised the study protocol, and approved the final manuscript. K.S. is responsible for the air pollution modelling, revised the study protocol, and approved the final manuscript. A.D. (Anna Degórska) is responsible for the air pollution modelling, revised the study protocol, and approved the final manuscript. J.B. is responsible for the air pollution modelling, revised the study protocol, and approved the final manuscript. B.K. is responsible for the quality control of the MRI data, revised the study protocol, and approved the final manuscript. A.D. (Aleksandra Domagalik) contributed to the design of the MRI acquisition protocol, revised the study protocol, and approved the final manuscript. M.S. is the principal investigator of the study and obtained the necessary funds, designed the study, in particular the behavioural and MRI acquisition protocol, conceptualized the fieldwork and is overseeing it, drafted the first version of the study protocol, revised the draft, and approved the final manuscript. All authors have read and agreed to the published version of the manuscript. The "NeuroSmog: Determining the impact of air pollution on the developing brain" (Nr. POIR.04.04.00-1763/18-00) project is implemented as part of the TEAM-NET programme of the Foundation for Polish Science, co-financed from EU resources obtained from the European Regional Development Fund under the Smart Growth Operational Programme. The aforementioned funding sources had no involvement in the design of the study, collection, analysis nor interpretation of the data, writing of the report, nor in the decision to submit the manuscript for publication. The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Ethical Committee of the Institute of Psychology, Jagiellonian University, Kraków, Poland (#KE_24042019A). The Clinical Trials Identifier is NCT04574414. Informed Consent Statement: Written informed consent was obtained from all children involved in the study and their legal guardians. The data collected in this study will be available on request from the correspondin author. The data are not publicly available due to the local ethical and legal restrictions. We are very grateful to all of the children and their parents for their participation in the study, partner schools for helping us to reach out to children assigned to the control group, field psychologists for identifying children with ADHD and performing psychological testing on all children, and the study team for technical, logistic, administrative, and communication efforts. We would also like to thank the project's Advisory Board members-Serena Counsell, Frank J. Kelly, Michał Krzyżanowski, Franck Ramus, Jordi Sunyer, Tomasz Ulanowski, and Wiktor Warchałowskifor their scientific support and advice. Finally, we would also like to thank Kees de Hoogh for assistance with air pollution modelling. The authors declare that they have no conflicts of interest. Environmental and Health Impacts of Air Pollution: A Review. Front. Public Health 2020, 8, 14 Preventing Disease through Health Environments: A Global Assessment of the Burden of Disease from Environmental Risks Estimates and 25-year trends of the global burden of disease attributable to ambient air pollution: An analysis of data from the Global Burden of Diseases Study Environmental determinants of cardiovascular disease: Lessons learned from air pollution Air Pollution and Cardiovascular Disease: JACC State-of-the-Art Review The influence of childhood traffic-related air pollution exposure on asthma, allergy and sensitization: A systematic review and a meta-analysis of birth cohort studies Early-life exposure to air pollution and childhood allergic diseases: An update on the link and its implications Long-term Fine Particulate Matter Exposure and Nonaccidental and Cause-specific Mortality in a Large National Cohort of Chinese Men Association of Solid Fuel Use with Risk of Cardiovascular and All-Cause Mortality in Rural China Acute effects of fine particulate matter constituents on mortality: A systematic review and meta-regression analysis Ambient black carbon particles reach the fetal side of human placenta Combined effects of ambient air pollution and home environmental factors on low birth weight Early life exposure to air pollution particulate matter (PM) as risk factor for attention deficit/hyperactivity disorder (ADHD): Need for novel strategies for mechanisms and causalities Alzheimer disease starts in childhood in polluted Metropolitan Mexico City. A major health crisis in progress Magnetite pollution nanoparticles in the human brain Fine particulate matter aggravates intestinal and brain injury and affects bacterial community structure of intestine and feces in Alzheimer's disease transgenic mice Fine particulate matter exposure during childhood relates to hemispheric-specific differences in brain structure Traffic-related particulate matter affects behavior, inflammation, and neural integrity in a developmental rodent model Proteomic changes driven by urban pollution suggest particulate matter as a deregulator of energy metabolism, mitochondrial activity, and oxidative pathways in the rat brain Concentrated ambient fine particulate matter (PM 2.5 ) exposure induce brain damage in pre and postnatal exposed mice Outdoor Air Pollution and Brain Structure and Function from Across Childhood to Young Adulthood: A Methodological Review of Air Pollution Exposure During Fetal Life, Brain Morphology, and Cognitive Function in School-Age Children Effects of prenatal exposure to particulate matter air pollution on corpus callosum and behavioral problems in children Traffic pollution exposure is associated with altered brain connectivity in school children Prenatal air pollution and childhood IQ: Preliminary evidence of effect modification by folate Association between ambient gaseous and particulate air pollutants and attention deficit hyperactivity disorder (ADHD) in children; a systematic review The air of Europe: Where are we going? Air quality status and trend in Europe Urban population exposure to air pollution in Europe over the last decades Premature mortality due to air pollution in European cities: A health impact assessment The Condition of Air Pollution in Krakow Ambient PM 2.5 Exposure and Mortality Due to Lung Cancer and Cardiopulmonary Diseases in Polish Cities Burden of Mortality and Disease Attributable to Multiple Air Pollutants in Warsaw, Poland Ozone, NO 2 and NOx Spatial Estimates and Their Uncertainties World Health Organization. International Classification of Diseases for Mortality and Morbidity Statistics (11th revision) R: A Language and Environment for Statistical Computing R Foundation for Statistical Computing Urban residential greenspace and mental health in youth: Different approaches to testing multiple pathways yield different conclusions Multiple pathways link urban green-and bluespace to mental health in young adults GINIplus and LISAplus-Design and selected results of two German birth cohorts about natural course of atopic diseases and their determinants Associations of Preconception Exposure to Air Pollution and Greenness with Offspring Asthma and Hay Fever Podręcznik Diagnosty Bateria diagnozy funkcji poznawczych PU1. Pamięć-uwagafunkcje wykonawcze. In Diagnostic Battery for Cognitive Functions Evaluation: Memory-Attention-Executive Functions Cognitive dimensions of Learning in Children with Problems in Attention, Learning, and Memory Conners 3-Questionnaires for ADHD Assessment Multicultural Understanding of Child and Adolescent Psychopathology: Implications for Mental Health Assessment Zaburzenia Emocjonalne i Behawioralne u Dzieci i Młodzieży Szkolnej w Polsce; Rozprawa Habilitacyjna The Achenbach System of Empirically Based Assessment (ASEBA): Development, Findings, Theory, and Applications The Polish Adaptation of FACES EV-SOR Development of attentional networks in childhood Testing the efficiency and independence of attentional networks Continuous Performance Test Evidence of difficulties in sustained attention in children with ADDH Continuous performance test performance in a normative epidemiological sample Reliability and validity of the Conners' Continuous Performance Test The Adolescent Brain Cognitive Development (ABCD) study: Imaging acquisition across 21 sites Evaluation of slice accelerations using multiband echo planar imaging at 3 T. Neuroimage Multiplexed echo planar imaging for sub-second whole brain FMRI and fast diffusion imaging Brain morphometry with multiecho MPRAGE Mapping human cortical areas in vivo based on myelin content as revealed by T1-and T2-weighted MRI Volumetric navigators for prospective motion correction and selective reacquisition in neuroanatomical MRI MP2RAGE, a self bias-field corrected sequence for improved segmentation and T1-mapping at high field Diffusion tensor imaging (DTI)-based white matter mapping in brain research: A review Opportunities and limitations of intrinsic functional connectivity MRI Modalities, modes, and models in functional neuroimaging What we can do and what we cannot do with fMRI The development of inhibitory control in preschool children: Effects of "executive skills" training Sex differences in child-onset, life-course-persistent conduct disorder: A review of biological influences Executive cognitive functions and impulsivity as correlates of risk taking and problem behavior in preadolescents The Lifespan Human Connectome Project in Development: A large-scale study of brain connectivity development in 5-21 year olds Optimal experimental design for event-related fMRI. Hum. Brain Mapp An fMRI Study of the Impact of Block Building and Board Games on Spatial Ability Development of land use regression models for particle composition in twenty study areas in Europe Spatial PM 2.5 , NO 2 , O 3 and BC models for Western Europe-Evaluation of spatiotemporal stability A comparison of linear regression, regularization, and machine learning algorithms to develop Europe-wide spatial models of fine particles and nitrogen dioxide Chief Inspectorate of Environmental Protection Emission projections and limit values of air pollution concentration-A case study using the EMEP4PL model EMEP4PL and WRF-Chem-Evaluation of the Modelling Results QGIS Development Team; QGIS Geographic Information System. Open Source Geospatial Foundation Project Robust causal inference using directed acyclic graphs: The R package 'dagitty'