Neurobiological, Psychosocial, and Behavioral Mechanisms Mediating Associations Between Physical Activity and Psychiatric Symptoms in Youth in the Netherlands

María Rodriguez-Ayllon; Alexander Neumann; Amy Hofman; Trudy Voortman; David R Lubans; Junwen Yang-Huang; Pauline W Jansen; Hein Raat; Meike W Vernooij; Ryan L Muetzel

doi:10.1001/jamapsychiatry.2023.0294

. 2023 Mar 29;80(5):451–458. doi: 10.1001/jamapsychiatry.2023.0294

Neurobiological, Psychosocial, and Behavioral Mechanisms Mediating Associations Between Physical Activity and Psychiatric Symptoms in Youth in the Netherlands

María Rodriguez-Ayllon ¹, Alexander Neumann ^2,³, Amy Hofman ¹, Trudy Voortman ^1,⁴, David R Lubans ^5,^6,⁷, Junwen Yang-Huang ⁸, Pauline W Jansen ^9,¹⁰, Hein Raat ⁶, Meike W Vernooij ^1,^11,^✉, Ryan L Muetzel ^9,¹¹

¹Department of Epidemiology, Erasmus University Medical Center Rotterdam, Rotterdam, the Netherlands

²Department of Biomedical Sciences, University of Antwerp, Antwerp, Belgium

³VIB Center for Molecular Neurology, Antwerp, Belgium

⁴Division of Human Nutrition and Health, Wageningen University & Research, Wageningen, the Netherlands

⁵Centre for Active Living and Learning, University of Newcastle, Callaghan, New South Wales, Australia

⁶Hunter Medical Research Institute, New Lambton Heights, New South Wales, Australia

⁷Faculty of Sport and Health Sciences, University of Jyväskylä, Jyväskylä, Finland

⁸Department of Public Health, Erasmus Medical Center, Rotterdam, the Netherlands

⁹Department of Child and Adolescent Psychiatry/Psychology, Erasmus University Medical Center Rotterdam, Rotterdam, the Netherlands

¹⁰Department of Psychology, Education and Child Studies, Erasmus University Medical Center Rotterdam, Rotterdam, the Netherlands

¹¹Department of Radiology and Nuclear Medicine, Erasmus University Medical Center, Rotterdam, the Netherlands

Accepted for Publication: January 23, 2023.

Published Online: March 29, 2023. doi:10.1001/jamapsychiatry.2023.0294

^✉

Corresponding Author: Meike W. Vernooij, MD, PhD, Department of Radiology and Nuclear Medicine, Erasmus University Medical Center Rotterdam, PO Box 2040, 3000 CA, Rotterdam, the Netherlands (m.vernooij@erasmusmc.nl).

Author Contributions: Drs Rodriguez-Ayllon and Muetzel had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Rodriguez-Ayllon, Hofman, Voortman, Raat, Vernooij, Muetzel.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Rodriguez-Ayllon, Jansen, Vernooij, Muetzel.

Critical revision of the manuscript for important intellectual content: Rodriguez-Ayllon, Neumann, Hofman, Voortman, Lubans, Yang-Huang, Raat, Muetzel.

Statistical analysis: Rodriguez-Ayllon, Neumann, Hofman, Vernooij.

Obtained funding: Muetzel.

Supervision: Voortman, Jansen, Raat, Vernooij, Muetzel.

Conflict of Interest Disclosures: None reported.

Funding/Support: Dr Rodriguez-Ayllon was supported by the Ramon Areces Foundation and the Alicia Koplowitz Foundation. This study was supported by the Sophia Foundation (S18-20) and Netherlands Organization for Health Research and Development (ZonMw). Supercomputing resources were supported by the Netherlands Organization for Scientific Research (Exacte Wetenschappen), and SURFsara (Snellius Compute Cluster, www.surfsara.nl). The Generation R Study is conducted by the Erasmus University Medical Center in close collaboration with the School of Law and the Faculty of Social Sciences of the Erasmus University Rotterdam, the Municipal Health Service Rotterdam area, and the Stichting Trombosedienst en Artsenlaboratorium Rijnmond. The general design of the Generation R Study was made possible by financial support from the Erasmus University Medical Center Rotterdam, ZonMw, the Netherlands Organization for Scientific Research, and the Ministry of Health, Welfare, and Sport. Dr Lubans is supported by a National Health and Medical Research Council research fellowship (APP1154507).

Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Data Sharing Statement: See Supplement 2.

Additional Contributions: We gratefully acknowledge the contributions of the participating children and parents, general practitioners, hospitals, midwives, and pharmacies in Rotterdam.

^✉

Corresponding author.

PMCID: PMC10061317 PMID: 36988919

Key Points

Question

Which neurobiological, psychosocial, or behavioral pathways mediate the associations between physical activity and psychiatric symptoms in young people?

Findings

This cohort study found that self-esteem mediated the association between sports participation in childhood and internalizing symptoms in adolescence.

Meaning

Physical activity interventions carried out during childhood should consider self-esteem improvements as a channel to protect young people against the later emergence of internalizing problems.

This cohort study evaluates neurobiological, psychosocial, and behavioral mechanisms potentially underlying associations between physical activity and psychiatric symptoms in children in the Netherlands.

Abstract

Importance

Understanding the mechanisms by which physical activity is associated with a lower risk of psychiatric symptoms may stimulate the identification of cost-efficient strategies for preventing and treating mental illness at early life stages.

Objective

To examine neurobiological, psychosocial, and behavioral mechanisms that mediate associations of physical activity with psychiatric symptoms in youth by testing an integrated model.

Design, setting, and participants

Generation R is an ongoing prospective population-based cohort study collecting data from fetal life until young adulthood in a multiethnic urban population in the Netherlands. Pregnant women living in Rotterdam with an expected delivery date between April 2002 and January 2006 were eligible for participation along with their children born during this time. Data were collected at a single research center in the Erasmus Medical Center Sophia Children’s Hospital. For the current study, data were analyzed from 4216 children with complete data on both exposure and outcome at ages 6, 10, and 13 years. Data were analyzed from January 2021 to November 2022.

Exposures

Physical activity was ascertained at age 6 years (visit 1) via parent report and included weekly frequency and duration of walking or cycling to or from school, physical education at school, outdoor play, swimming, and sports participation.

Main Outcomes and Measures

Psychiatric symptoms (internalizing and externalizing symptoms) were assessed at age 6 years (visit 1) and at age 13 years (visit 3) using the Child Behavior Checklist. Several mechanisms were explored as mediators, measured at age 10 years (visit 2). Neurobiological mechanisms included total brain volume, white matter microstructure, and resting-state connectivity assessed using a 3-T magnetic resonance imaging scanner. Psychosocial mechanisms included self-esteem, body image, and friendship. Behavioral mechanisms included sleep quality, diet quality, and recreational screen time. Pearson correlations between physical activity measures and psychiatric symptoms were calculated, with false discovery rate correction applied to account for the number of tests performed. Mediation analyses were performed when a correlation (defined as false discovery rate P < .05) between exposure and outcome was observed and were adjusted for confounders.

Results

Among the 4216 children included in this study, the mean (SD) age was 6.0 (0.4) years at visit 1, and 2115 participants (50.2%) were girls. More sports participation was associated with fewer internalizing symptoms (β for direct effect, −0.025; SE, 0.078; P = .03) but not externalizing symptoms. Self-esteem mediated the association between sports participation and internalizing symptoms (β for indirect effect, −0.009; SE, 0.018; P = .002). No evidence was found for associations between any other neurobiological, psychosocial, or behavioral variables. No association was found between other types of physical activity and psychiatric symptoms at these ages.

Conclusions and Relevance

The integrated model presented in this cohort study evaluated potential mechanisms mediating associations between physical activity and psychiatric symptoms in youth. Self-esteem mediated an association between sports participation in childhood and internalizing symptoms in adolescence; other significant mediations were not observed. Further studies might explore whether larger effects are present in certain subgroups (eg, children at high risk of developing psychiatric symptoms), different ages, or structured sport-based physical activity interventions.

Introduction

The transition from childhood to adolescence involves extensive developmental changes that coincide with increased vulnerability to psychiatric symptoms.¹ Risk factors for psychiatric symptoms have been well established.¹ However, less is known about protective factors for psychiatric symptoms in youth.

Compelling evidence demonstrated that physical activity positively affects mental health from childhood to adulthood.^2,3,4,5 The strength of the evidence has led the World Health Organization to include psychiatric symptoms, such as depression and anxiety, among the conditions that can be prevented through physical activity in their most recent guidelines.^4,6 However, the pathways between physical activity and mental health and the life stage at which these come into play remain uncertain. In this context, Lubans et al⁷ suggested a conceptual model that postulates 3 broad categories of mechanisms through which physical activity may be associated with mental health: neurobiological, psychosocial, and behavioral mechanisms.

The neurobiological mechanism hypothesis suggests that physical activity may alter brain structure or function and in turn reduce the development of psychiatric symptoms.^8,9 For instance, depression has been associated with a lower density of neuronal cells in the hippocampus,^10,11 a region that has shown structural plasticity in response to physical activity.⁸ Higher physical activity has also been associated with better white matter microstructure during childhood.¹² Nevertheless, it is unknown whether changes in white matter mediate the effect of physical activity on psychiatric symptoms.

The psychosocial mechanism theory proposes that physical activity might satisfy basic psychological needs, such as social connectedness, which in turn could decrease the risk of developing psychiatric symptoms.¹³ Extensive research has also shown the effect of physical activity on psychiatric symptoms is partially mediated by changes in self-perception.¹⁴

Improvements in psychiatric symptoms resulting from physical activity could be also mediated by changes in associated behaviors. These may include improved sleep, healthier eating habits, or reduced recreational screen time.¹⁵

Overall, some isolated mechanisms through which physical activity may reduce psychiatric symptoms have been identified.^13,16 Nevertheless, an integrated model examining the joint and independent contributions of the proposed mechanisms is lacking, making it difficult to obtain a comprehensive picture. We hypothesized that the association of physical activity with psychiatric symptoms operates via multiple mechanisms. Therefore, the aim of our study was to identify key mechanisms that may underly the association between physical activity and psychiatric symptoms in youth using an integrated perspective.¹⁷

Methods

Study Design and Participants

This study was part of the Generation R study, an ongoing prospective population-based birth cohort conducted in Rotterdam, the Netherlands. The full design is detailed elsewhere.^18,19 Briefly, around 10 000 pregnant women from the general population were enrolled in the study between April 2002 and January 2006, and data have been collected from them and their children over the past 20 years.^18,19 The current study used data from 3 visits, when the children were aged 6, 10, and 13 years. The medical ethics committee of Erasmus Medical Centre approved all study procedures. All participants provided written informed consent. The study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline (eAppendix in Supplement 1).

Sample

At the age of 6 years, 6265 participants provided physical activity data. Of these, 4216 participants provided data on psychiatric symptoms as well at the age of 13 years and thus had complete data on both exposure and outcome (eFigure 1 in Supplement 1).

Study Variables

Physical Activity

Physical activity was reported by the primary caregiver (97% mothers; other caregivers included fathers and other legal guardians). The questionnaire included frequency and duration that a child engaged in physical education at school, walking or cycling to or from school, outdoor play, swimming, and sports (ie, athletics, basketball, combined sports, dance, football, gymnastics, hockey, martial arts, tennis, others).¹² Time spent on each activity was calculated as days per week × hours per day). Total physical activity was calculated by adding the hours of active commuting, physical education at school, outdoor play, swimming, and sport participation.

Psychiatric Symptoms

Primary caregivers filled out the validated Child Behavior Checklist to report on children's psychiatric symptoms.^20,21 We examined the Child Behavior Checklist broadband subscales of internalizing problems (ie, depression, anxiety, and somatic symptoms) and externalizing problems (ie, conduct symptoms, rule-breaking behavior, and attention-deficit/hyperactivity disorder) as well as the 6 Syndrome Scale subdomains.

Neurobiological Mediators

High-resolution structural magnetic resonance imaging (MRI), diffusion-weighted white matter imaging, and resting-state functional MRI were collected on a 3-T MRI scanner.²² Structural MRI data were processed through FreeSurfer version 6.0.0 (Martinos Center for Biomedical Imaging), which yielded anatomical labels for broad tissue classes (eg, white and gray matter) and several brain structures (eg, hippocampus). Diffusion image preprocessing was conducted using the FMRIB Software Library version 5.0.9 (FMRIB Analysis Group).²³ Two metrics of white matter microstructure (ie, fractional anisotropy and mean diffusivity) were derived globally (ie, across multiple tracts) and for corpus callosum fibers (forceps major and minor). Dynamic functional network connectivity was estimated using the Group ICA of fMRI Toolbox GroupICAT version 4.0b (Matlab R2020a).²⁴

Psychosocial Mediators

This study analyzed self-esteem (ie, an individual’s evaluation of their qualities and limitations²⁵) using an 18-item questionnaire adapted from the Harter Self-Perception Profile for Children,²⁶ body image (ie, perceived physical attractiveness) using the Development of the Children’s Body Image scale,²⁷ and friendships (ie, a state of mutual trust and support between people) using an adapted version of the Parker and Asher Friendship Quality Questionnaire.^28,29

Behavioral Mediators

Sleep quality was evaluated using the Sleep Disturbance Scale for Children.^30,31 Higher scores indicate lower sleep quality. Diet quality was quantified by a predefined food-based diet quality score based on Dutch dietary recommendations for 8-year-old children.³² Recreational screen time was obtained through a parent-reported questionnaire.¹⁹ See the eMethods in Supplement 1 for further details on all study variables.

Confounders

Maternal education and national origin and child age, sex, body mass index (BMI; calculated as weight in kilograms divided by height in meters squared), and nonverbal IQ were included as confounders. Parental national origin was based on the country of birth of the mother and mother’s parents and was ascertained via questionnaire with categories conforming to those used by the Dutch Government Office for Statistics. Maternal education was defined by the highest completed education and divided into 2 categories ranging from low (from no education to high school or vocational training) to high education level (from higher vocational education to university). Child height and weight were measured at the research center and BMI was calculated and standardized according to the Dutch reference growth curves.^33,34 Nonverbal IQ was assessed using the Snijders-Oomen Niet-verbale intelligentie Test—Revisie version 2.5-7 (WorldCat).

Statistical Analyses

Statistical analyses were performed using R version 4.0.5 (R Foundation). First, we explored the Pearson correlation between physical activity measures and psychiatric symptoms, with false discovery rate correction applied to account for the number of tests performed.³⁵ Second, mediation analyses were performed with the Lavaan package version 0.6-9 (R Foundation) when a correlation (defined as false discovery rate P < .05) between exposure and outcome was observed. Physical activity was entered into the model as the exposure, and the neurobiological, psychosocial, and behavioral mechanisms were entered as mediators. Specifically, we explored the individual effect of each mediator on the association between physical activity and psychiatric symptoms. Additionally, we grouped individual mediators into 3 categories (ie, neurobiological, psychosocial, and behavioral) according to a previously proposed conceptual model⁷ and explored its summed indirect effect on the association between physical activity and psychiatric symptoms. Psychiatric symptoms indexed as broadband scales of internalizing problems and externalizing symptoms were entered as outcomes. An illustration of the general modeling strategy is depicted in eFigure 2 in Supplement 1. Mediation models were adjusted for several potential confounders, including baseline psychiatric symptoms at age 6; maternal education level and national origin; and child sex, age at visit 1, BMI, and IQ. Additionally, we tested whether the mediation mechanisms differed between girls and boys, children with different maternal education, and children with different BMI measures by performing mediation invariance analyses (multigroup analyses). A number of supplemental and sensitivity analyses were also run (eMethods in Supplement 1).

Maximum likelihood with robust standard errors was used to fit the structural equation models while accounting for missing data in mediators and confounders (full-information maximum likelihood), as implemented in Lavaan. This is a standard approach to prevent listwise deletion of participants with missing data.

Results

Sample Characteristics

The mean (SD) age of the study population was 6.0 (0.4) years at baseline, 9.8 (0.3) at visit 2, and 13.5 (0.4) at visit 3 (Table), and 2115 participants (50.2%) were girls. Characteristics of participants with complete cases in predictors, outcomes, and mediators are shown in eTable 1 in Supplement 1. Descriptive information on exposures, mediators, and outcomes is presented in eTable 2 in Supplement 1. At baseline, participants reported a total mean (SD) physical activity of 14.6 (8.1) hours per week. Compared to sports participation (mean [SD], 0.6 [0.8] hours per week), the levels of outdoor play were relatively high (mean [SD],11.2 [7.9] hours per week). Nonresponse information to ascertain how similar the study sample was to the original cohort is shown in eTable 3 in Supplement 1.

Table. Sample Characteristics (N = 4216).

Characteristic	Mean (SD)
Child characteristics
Female, no. (%)^a	2115 (50.2)
Male, no. (%)^a	2101 (49.8)
Age at visit 1, y	6.0 (0.4)
Age at visit 2, y	9.8 (0.3)
Age at visit 3, y	13.5 (0.4)
Body mass index^b	16.0 (1.6)
Behavior problems, sum score (CBCL)	18.7 (15.3)
Nonverbal IQ	103.4 (14.6)
Parental characteristics
Maternal education, no. (%)^a^,^c
Higher education	2670 (63.4)
Lower education	1543 (36.6)
National origin, no. (%)^a
Dutch	2849 (68.0)
Other	1363 (32.0)

Open in a new tab

Abbreviation: CBCL, Child Behavior Checklist school age.

^{^a}

Numbers do not add to 100% given missing data on sex for 1 participant, maternal education for 3 participants, and national origin for 4 participants.

^{^b}

Calculated as weight in kilograms divided by height in meters squared

^{^c}

Maternal education level was defined by the highest completed education and divided into 2 categories ranging from low (from no education to high school) to high (from higher vocational education to university).

Correlation Between Physical Activity and Psychiatric Symptoms

A correlation matrix of physical activity and psychiatric symptoms is presented in Figure 1. Higher levels of sports participation at age 6 years were correlated with lower levels of internalizing symptoms at age 13 years (r, −0.063; adjusted P = .001). No other correlations were observed for other measures of physical activity and psychiatric symptoms. Therefore, mediation analyses were only carried out with sports participation as the predictor and internalizing symptoms as the outcome (eFigure 2 in Supplement 1).

Mediation Analyses

The results of the overall integrative mediation model are presented in Figure 2. Higher levels of sports participation were associated with lower internalizing symptoms (β for direct effect, −0.025; SE, 0.078; P = .03). Only self-esteem was found to mediate the association between sports participation and internalizing symptoms (β for indirect effect, −0.009; SE, 0.018; P = .002). Specifically, self-esteem explained 26% of the variance (calculated as β for indirect effect / β for total effect) in the association between sports participation and internalizing symptoms. Independently, higher levels of sports participation were associated with higher self-esteem (β for direct effect, 0.059; SE, 0.084; P < .001), and higher self-esteem was associated with lower internalizing symptoms (β for direct effect, −0.146; SE, 0.027; P < .001). In a post hoc exploratory analysis, we detected the mediating role of self-esteem was mainly driven by the athletic competence domain (eFigure 3 in Supplement 1).

Independently, higher sports participation was associated with a better diet quality (β for direct effect, 0.049; SE, 0.028; P = .01), while lower sleep quality was associated with higher internalizing symptoms (β for direct effect, 0.082; SE, 0.041; P < .001).

Multigroup analyses showed no differences between girls and boys (χ² = 72.06; P = .18), and between children with different BMIs (χ² = 134.63; P = .24). In contrast, we found differences between children from families with lower vs higher educational status (χ² = 115.54; P < .001). In our stratified analyses, self-esteem mediated the association of sports participation with internalizing problems among those with lower levels of maternal education (β for indirect effect, −0.021; SE, 0.044; P = .001) but not among those with higher levels of maternal education (β for indirect effect, −0.004; SE, 0.020; P = .20). See eFigures 4 and 5 in Supplement 1 for further details. Several additional supplemental analyses were run to examine the specificity and sensitivity of the results (eg, specific psychiatric symptoms) and are presented in the eResults section and eTables 6-14 in Supplement 1.

Discussion

In this study, we sought to shed new light onto the association between physical activity and mental health in youth. Specifically, using in-depth neurobiological, psychological, and behavioral measures gathered from a large representative sample of more than 4000 youth, we observed that self-esteem mediated the association between sports and internalizing symptoms. Thus, more participation in sports was associated with increased self-esteem, which in turn was associated with lower levels of internalizing problems at follow-up, independent of baseline mental health status. This finding was particularly relevant in children with caregivers with low educational attainment.

Sports participation was inversely associated with internalizing symptoms in youth. However, this association was relatively small. Most associations in this study are in line with previous studies.^{5,36,37,38,39} For instance, involvement in sports during childhood was negatively associated with depressive symptoms in young adulthood; however, the association was small, especially after including potential confounders.³⁹ Additionally, research has argued that in trials with controlled clinical samples, physical activity has shown a larger and more beneficial effect on psychiatric symptoms compared to studies involving the general population.³⁷ Taken together, these findings suggest that larger effect sizes might be observed in studies including clinical samples of adolescents diagnosed with major psychological disorders.^36,37,38 We did not observe associations between other types of physical activities and psychiatric symptoms, which suggests that practicing sports during early childhood might be the most effective physical activity practice to improve or preserve mental health among adolescents.

Self-esteem mediated the association between sports and internalizing symptoms in youth. Adolescents have been observed to shape their self-esteem by developing skills, discovering preferences, and associating themselves with others.⁴⁰ Sports activities offer youth a means to develop their self-esteem, distinguish themselves from others, and operate in a challenging setting outside of academics.⁴⁰ Therefore, it is possible that early participation in sports could provide children with greater maturity during adolescence, which might help them deal with new life circumstances (eg, academic pressure and the influence of peers) and protect their mental health. This finding is consistent with a recent systematic review⁴¹ wherein self-dimensions were the only consistent paths observed through which physical activity was associated with improvement in psychiatric symptoms in youth.

The mediating role of self-esteem was mainly driven by the athletic competence domain, referring to one’s ability to do well at sports. These results further support the idea that youth with high perceived competence in sports are more likely to experience the positive effects of sports on mental health.⁴² As a result, future sports-based interventions designed to protect young people’s mental health might consider the use of evidence based-physical activity strategies (eg, the supportive, active, autonomous, fair, enjoyable [SAAFE] principles⁴³) as well as young people’s sports preferences.⁴⁴

Self-esteem mediated the association between sports participation and internalizing symptoms, particularly among children of caregivers with low education. Home environments, low maternal education, or low socioeconomic status can act as early life adversities in the context of emerging psychiatric symptoms in childhood.^45,46,47 However, some children in the same circumstances may be more resilient to the development of psychiatric symptoms. This fact could be partially explained by the interaction of intrapersonal resilience factors, such as IQ, self-identity, or self-esteem.⁴⁷ Specifically, self-esteem has been identified as a potential mediator in the association between early life adversities and the development of psychiatric symptoms.⁴⁸ Self-esteem might be improved by effective sport-based interventions.^49,50 Therefore, future studies should explore whether improving self-esteem through early sports-based interventions may protect the overall mental health of youth exposed to early life adversities.

We did not observe any other mechanisms underlying the association of sports participation with internalizing symptoms. Accordingly, the aforementioned systematic review⁴¹ showed the role of the neurobiological mechanisms in the association between physical activity and psychiatric symptoms is unclear, probably because of the inconsistencies and heterogeneity observed among included studies. For instance, previous studies have used MRI data as indicators of the neurobiological mechanisms while others have examined the role of blood circulating biomarkers.⁴¹ In healthy young individuals, neurobiological measurements in the form of circulating blood biomarkers might offer a more dynamic indication of the role of neurobiological mechanisms in the association between physical activity and psychiatric symptoms. Specifically, a 20-week physical exercise intervention was associated with a reduction in circulating macrophage scavenger receptor type I levels in children.⁵¹ Macrophage scavenger receptor type I is a membrane glycoprotein expressed in macrophages and has been associated with neurobiological processes and neurological diseases.⁵² In contrast, the same intervention was not associated with the structural and functional brain outcomes explored in another study.⁵³ Other potential reasons, such as neurodevelopmental differences between the children, the need for more advanced imaging methods, or whole-brain vs region-specific approaches, could be clouding the potential role of neurobiological mediators in this association. Future studies should explore other psychosocial (eg, enjoyment) and behavioral (eg, coping skills) mechanisms.

We used data across 3 time points from 1 of the largest cohorts of youth with information on physical activity, behavioral and emotional measures, and neuroimaging worldwide. A strength of this study was the unique inclusion of a broad set of mechanisms into a previously described integrated model that allowed us to obtain an overall picture of the mechanisms mediating physical activity and psychiatry symptoms in youth. Further, prospectively collected data across different points in time allowed us to model these mechanisms within a mediation framework.

Limitations

Our findings must be interpreted in the context of relevant limitations. First, the observational design limits inferences about causality to any of the associated factors, and residual confounding cannot be ruled out. Second, other potential mechanisms not included in the model could also mediate the association between sports and psychiatry symptoms. Third, we studied the mechanisms underlying the long-term associations of sports with psychiatry symptoms from childhood to adolescence. It is possible that more immediate effects of sports on psychopathology, (eg, within days or months) act via different mediators, which could be explored in future research using more high-frequency repeated measures. Fourth, we measured the predictor and mediators at a single time point, which did not allow us to explore the stability of those variables from childhood to adolescence. Similarly, the precise reliability of some self-reported mediators at such a young age remains unclear. Fifth, physical activity was assessed by caregiver reports, leading to the possibility of underestimation or overestimations of the behaviors. Additionally, both the predictor and the outcome were reported by the primary caregivers, which could overestimate the association observed due to shared method variance. However, sensitivity analyses showed that using the child as the reporter at the outcome did not change the overall results. Sixth, despite being multiethnic and diverse, the study sample available for analysis consisted of, for example, more individuals of European descent and more highly educated individuals when compared to the original sample at study enrollment.

Conclusions

In this cohort study, sports participation during early childhood was modestly associated with internalizing symptoms in adolescence. We did not observe associations between other types of physical activities and psychiatric symptoms at these ages. Among all neurobiological, psychological, and behavioral information examined, self-esteem was identified as a mediating factor through which sports and internalizing problems in youth were associated. Further studies might explore whether larger effects are present in certain subgroups (eg, children at high risk of developing psychiatric symptoms), different ages, or structured sport-based interventions.

Supplement 1.

eAppendix. STROBE Statement—Checklist of items that should be included in reports of cohort studies

eMethods

eResults

eTable 1. Sample characteristics of participants with complete cases in predictors, outcomes, and mediators (n=1,025)

eTable 2. Exposures, mediators, and outcomes characteristics

eTable 3. Included-excluded comparison sample characteristics

eFigure 1. Flow chart

eFigure 2. Visualization of the mediation modeling approach

eFigure 3. Integrative mediation model to explore the mediating role of self-esteem domains in the relationship between sports participation and internalizing symptoms in young people (n=4,216): a posteriori analysis

eFigure 4. Integrative mediation model on the mechanisms linking sports participation and internalizing symptoms in children whose caregivers have a low educational status (n=1,543)

eFigure 5. Integrative mediation model on the mechanisms linking sports participation and internalizing symptoms in children whose caregivers have a high educational status (n=2,670)

eFigure 6. Integrative mediation model on the mechanisms linking sports participation and internalizing symptoms in young people (n=4,216) including hippocampal volume and corpus callosum fractional anisotropy instead of global measures

eFigure 7. Integrative mediation model on the mechanisms linking sports participation and somatic complaints in young people (n=4,216)

eFigure 8. Integrative mediation model on the mechanisms linking sports participation and anxious/depressed mood in young people (n=4,216)

eFigure 9. Integrative mediation model on the mechanisms linking sports participation and withdrawn/depressed mood in young people (n=4,216)

eFigure 10. Integrative mediation model on the mechanisms linking sports participation and internalizing symptoms in young people including only participants with complete data (n=1,025)

eFigure 11. Integrative mediation model on the mechanisms linking sports participation and internalizing symptoms in young people excluding siblings randomly (n=3,921)

eFigure 12. Integrative mediation model, without adjusting for internalizing symptoms at the baseline, on the mechanisms linking sports participation and internalizing symptoms in young people (n=4216)

eFigure 13. Integrative mediation model on the mechanisms linking sports participation and internalizing symptoms in young people (n=4,060) with internalizing symptoms reported by the children instead of the primary caregiver

eFigure 14. Mediation model including individually the mechanisms linking sports participation and internalizing symptoms in young people (n=4,060)

Click here for additional data file.^{(2.9MB, pdf)}

Supplement 2.

Data sharing statement

Click here for additional data file.^{(16.2KB, pdf)}

References

1.Patel V, Flisher AJ, Hetrick S, McGorry P. Mental health of young people: a global public-health challenge. Lancet. 2007;369(9569):1302-1313. doi: 10.1016/S0140-6736(07)60368-7 [DOI] [PubMed] [Google Scholar]
2.Rodriguez-Ayllon M, Estévez-López F, Cadenas-Sanchez C, et al. Physical activity, sedentary behaviour and mental health in young people: a review of reviews. In: Pingitore A, Mastorci F, Vassalle C, eds. Adolescent Health and Wellbeing: Current Strategies and Future Trends. Springer International Publishing; 2019:35-73. doi: 10.1007/978-3-030-25816-0_3 [DOI] [Google Scholar]
3.Rodriguez-Ayllon M, Cadenas-Sánchez C, Estévez-López F, et al. Role of physical activity and sedentary behavior in the mental health of preschoolers, children and adolescents: a systematic review and meta-analysis. Sports Med. 2019;49(9):1383-1410. doi: 10.1007/s40279-019-01099-5 [DOI] [PubMed] [Google Scholar]
4.Chaput JP, Willumsen J, Bull F, et al. 2020 WHO guidelines on physical activity and sedentary behaviour for children and adolescents aged 5-17 years: summary of the evidence. Int J Behav Nutr Phys Act. 2020;17(1):141. doi: 10.1186/s12966-020-01037-z [DOI] [PMC free article] [PubMed] [Google Scholar]
5.Choi KW, Chen CY, Stein MB, et al. ; Major Depressive Disorder Working Group of the Psychiatric Genomics Consortium . Assessment of bidirectional relationships between physical activity and depression among adults: a 2-sample mendelian randomization study. JAMA Psychiatry. 2019;76(4):399-408. doi: 10.1001/jamapsychiatry.2018.4175 [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Bull FC, Al-Ansari SS, Biddle S, et al. World Health Organization 2020 guidelines on physical activity and sedentary behaviour. Br J Sports Med. 2020;54(24):1451-1462. doi: 10.1136/bjsports-2020-102955 [DOI] [PMC free article] [PubMed] [Google Scholar]
7.Lubans D, Richards J, Hillman C, et al. Physical activity for cognitive and mental health in youth: a systematic review of mechanisms. Pediatrics. 2016;138(3):e20161642-e20161642. doi: 10.1542/peds.2016-1642 [DOI] [PubMed] [Google Scholar]
8.Stillman CM, Esteban-Cornejo I, Brown B, Bender CM, Erickson KI. Effects of exercise on brain and cognition across age groups and health states. Trends Neurosci. 2020;43(7):533-543. doi: 10.1016/j.tins.2020.04.010 [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Erickson KI, Hillman C, Stillman CM, et al. ; FOR 2018 PHYSICAL ACTIVITY GUIDELINES ADVISORY COMMITTEE* . Physical activity, cognition, and brain outcomes: a review of the 2018 Physical Activity Guidelines. Med Sci Sports Exerc. 2019;51(6):1242-1251. doi: 10.1249/MSS.0000000000001936 [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Zhang FF, Peng W, Sweeney JA, Jia ZY, Gong QY. Brain structure alterations in depression: psychoradiological evidence. CNS Neurosci Ther. 2018;24(11):994-1003. doi: 10.1111/cns.12835 [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Valkenborghs SR, Hillman CH, Al-Iedani O, et al. Effect of high-intensity interval training on hippocampal metabolism in older adolescents. Psychophysiology. 2022;59(11):e14090. doi: 10.1111/psyp.14090 [DOI] [PMC free article] [PubMed] [Google Scholar]
12.Rodriguez-Ayllon M, Derks IPM, van den Dries MA, et al. Associations of physical activity and screen time with white matter microstructure in children from the general population. Neuroimage. 2020;205:116258. doi: 10.1016/j.neuroimage.2019.116258 [DOI] [PubMed] [Google Scholar]
13.Conley MI, Hindley I, Baskin-Sommers A, Gee DG, Casey BJ, Rosenberg MD. The importance of social factors in the association between physical activity and depression in children. Child Adolesc Psychiatry Ment Health. 2020;14(1):28. doi: 10.1186/s13034-020-00335-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Petty KH, Davis CL, Tkacz J, Young-Hyman D, Waller JL. Exercise effects on depressive symptoms and self-worth in overweight children: a randomized controlled trial. J Pediatr Psychol. 2009;34(9):929-939. doi: 10.1093/jpepsy/jsp007 [DOI] [PMC free article] [PubMed] [Google Scholar]
15.Lang C, Brand S, Feldmeth AK, Holsboer-Trachsler E, Pühse U, Gerber M. Increased self-reported and objectively assessed physical activity predict sleep quality among adolescents. Physiol Behav. 2013;120:46-53. doi: 10.1016/j.physbeh.2013.07.001 [DOI] [PubMed] [Google Scholar]
16.Gorham LS, Jernigan T, Hudziak J, Barch DM. Involvement in sports, hippocampal volume, and depressive symptoms in children. Biol Psychiatry Cogn Neurosci Neuroimaging. 2019;4(5):484-492. doi: 10.1016/j.bpsc.2019.01.011 [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Cerin E. Ways of unraveling how and why physical activity influences mental health through statistical mediation analyses. Ment Health Phys Act. 2010;3(2):51-60. doi: 10.1016/j.mhpa.2010.06.002 [DOI] [Google Scholar]
18.Jaddoe VWV, Mackenbach JP, Moll HA, et al. The Generation R study: design and cohort profile. Eur J Epidemiol. 2006;21(6):475-484. doi: 10.1007/s10654-006-9022-0 [DOI] [PubMed] [Google Scholar]
19.Kooijman MN, Kruithof CJ, van Duijn CM, et al. The Generation R Study: design and cohort update 2017. Eur J Epidemiol. 2016;31(12):1243-1264. doi: 10.1007/s10654-016-0224-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
20.Achenback TM, Rescorla LA. Manual for the ASEBA School-Age Forms and Profiles. University of Vermont, Center for Children, Youth and Families; 2001. [Google Scholar]
21.Ivanova MY, Achenbach TM, Rescorla LA, et al. Preschool psychopathology reported by parents in 23 societies: testing the seven-syndrome model of the child behavior checklist for ages 1.5-5. J Am Acad Child Adolesc Psychiatry. 2010;49(12):1215-1224. doi: 10.1097/00004583-201012000-00007 [DOI] [PMC free article] [PubMed] [Google Scholar]
22.White T, Muetzel RL, El Marroun H, et al. Paediatric population neuroimaging and the Generation R study: the second wave. Eur J Epidemiol. 2018;33(1):99-125. doi: 10.1007/s10654-017-0319-y [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Jenkinson M, Beckmann CF, Behrens TEJ, Woolrich MW, Smith SM. FSL. Neuroimage. 2012;62(2):782-790. doi: 10.1016/j.neuroimage.2011.09.015 [DOI] [PubMed] [Google Scholar]
24.López-Vicente M, Agcaoglu O, Pérez-Crespo L, et al. Developmental changes in dynamic functional connectivity from childhood into adolescence. Front Syst Neurosci. 2021;15:724805. doi: 10.3389/fnsys.2021.724805 [DOI] [PMC free article] [PubMed] [Google Scholar]
25.Sonstroem RJ, Morgan WP. Exercise and self-esteem: rationale and model. Med Sci Sports Exerc. 1989;21(3):329-337. doi: 10.1249/00005768-198906000-00018 [DOI] [PubMed] [Google Scholar]
26.Harter S. Manual for the Self-Perception Profile for Children. University of Denver; 1985. [Google Scholar]
27.Truby H, Paxton SJ. Development of the Children’s Body Image Scale. Br J Clin Psychol. 2002;41(Pt 2):185-203. doi: 10.1348/014466502163967 [DOI] [PubMed] [Google Scholar]
28.Parker JG, Asher SR. Friendship and friendship quality in middle childhood: links with peer group acceptance and feelings of loneliness and social dissatisfaction. Dev Psychol. 1993;29(4):611-621. doi: 10.1037/0012-1649.29.4.611 [DOI] [Google Scholar]
29.de Lijster JM, van den Dries MA, van der Ende J, et al. Developmental trajectories of anxiety and depression symptoms from early to middle childhood: a population-based cohort study in the Netherlands. J Abnorm Child Psychol. 2019;47(11):1785-1798. doi: 10.1007/s10802-019-00550-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
30.Bruni O, Ottaviano S, Guidetti V, et al. The Sleep Disturbance Scale for Children (SDSC). construction and validation of an instrument to evaluate sleep disturbances in childhood and adolescence. J Sleep Res. 1996;5(4):251-261. doi: 10.1111/j.1365-2869.1996.00251.x [DOI] [PubMed] [Google Scholar]
31.Koopman-Verhoeff ME, Serdarevic F, Kocevska D, et al. Preschool family irregularity and the development of sleep problems in childhood: a longitudinal study. J Child Psychol Psychiatry. 2019;60(8):857-865. doi: 10.1111/jcpp.13060 [DOI] [PMC free article] [PubMed] [Google Scholar]
32.van der Velde LA, Nguyen AN, Schoufour JD, et al. Diet quality in childhood: the Generation R Study. Eur J Nutr. 2019;58(3):1259-1269. doi: 10.1007/s00394-018-1651-z [DOI] [PMC free article] [PubMed] [Google Scholar]
33.Fredriks AM, van Buuren S, Burgmeijer RJ, et al. Continuing positive secular growth change in the Netherlands 1955-1997. Pediatr Res. 2000;47(3):316-323. doi: 10.1203/00006450-200003000-00006 [DOI] [PubMed] [Google Scholar]
34.Dutch Growth Research Foundation . Growth analyser. Accessed November 1, 2022. https://growthanalyser.org
35.Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc Series B Stat Methodol J R STAT SOC B. 1995;57(1):289-300. https://www.jstor.org/stable/2346101?seq=1#metadata_info_tab_contents [Google Scholar]
36.Carter T, Morres ID, Meade O, Callaghan P. The effect of exercise on depressive symptoms in adolescents: a systematic review and meta-analysis. J Am Acad Child Adolesc Psychiatry. 2016;55(7):580-590. doi: 10.1016/j.jaac.2016.04.016 [DOI] [PubMed] [Google Scholar]
37.Larun L. Nordheim LV, Ekeland E, Hagen KB, Heian F. Exercise in prevention and treatment of anxiety and depression among children and young people. Cochrane Database Syst Rev. 2006;(3). doi: 10.1002/14651858.CD004691.pub2 [DOI] [PMC free article] [PubMed] [Google Scholar]
38.Brown HE, Pearson N, Braithwaite RE, Brown WJ, Biddle SJH. Physical activity interventions and depression in children and adolescents: a systematic review and meta-analysis. Sports Med. 2013;43(3):195-206. doi: 10.1007/s40279-012-0015-8 [DOI] [PubMed] [Google Scholar]
39.Brunet J, Sabiston CM, Chaiton M, et al. The association between past and current physical activity and depressive symptoms in young adults: a 10-year prospective study. Ann Epidemiol. 2013;23(1):25-30. doi: 10.1016/j.annepidem.2012.10.006 [DOI] [PubMed] [Google Scholar]
40.Feldman AF, Matjasko JL. The role of school-based extracurricular activities in adolescent development: a comprehensive review and future directions. Rev Educ Res. 2016;75(2):159-210. doi: 10.3102/00346543075002 [DOI] [Google Scholar]
41.Nguyen Ho PT, Ha Pham Bich T, Tong T, et al. Mechanisms linking physical activity with psychiatric symptoms across the lifespan: A systematic review. medRxiv. doi: 10.1101/2022.07.13.22277479 [DOI] [PMC free article] [PubMed]
42.Kipp LE, Weiss MR. Physical activity and self-perceptions among children and adolescents. In: Ekkekakis P, ed. Routledge Handbook of Physical Activity and Mental Health. Routledge; 2013:187-199, https://experts.umn.edu/en/publications/physical-activity-and-self-perceptions-among-children-and-adolesc. [Google Scholar]
43.Lubans DR, Lonsdale C, Cohen K, et al. Framework for the design and delivery of organized physical activity sessions for children and adolescents: rationale and description of the ‘SAAFE’ teaching principles. Int J Behav Nutr Phys Act. 2017;14(1):24. doi: 10.1186/s12966-017-0479-x [DOI] [PMC free article] [PubMed] [Google Scholar]
44.Hulteen RM, Smith JJ, Morgan PJ, et al. Global participation in sport and leisure-time physical activities: a systematic review and meta-analysis. Prev Med. 2017;95:14-25. doi: 10.1016/j.ypmed.2016.11.027 [DOI] [PubMed] [Google Scholar]
45.Rijlaarsdam J, Stevens GWJM, van der Ende J, et al. Economic disadvantage and young children’s emotional and behavioral problems: mechanisms of risk. J Abnorm Child Psychol. 2013;41(1):125-137. doi: 10.1007/s10802-012-9655-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
46.Appleton AA, Buka SL, McCormick MC, Koenen KC, Loucks EB, Kubzansky LD. The association between childhood emotional functioning and adulthood inflammation is modified by early-life socioeconomic status. Health Psychol. 2012;31(4):413-422. doi: 10.1037/a0027300 [DOI] [PMC free article] [PubMed] [Google Scholar]
47.Zinn ME, Huntley ED, Keating DP. Resilience in adolescence: prospective self moderates the association of early life adversity with externalizing problems. J Adolesc. 2020;81:61-72. doi: 10.1016/j.adolescence.2020.04.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
48.Kim Y, Lee H, Park A. Patterns of adverse childhood experiences and depressive symptoms: self-esteem as a mediating mechanism. Soc Psychiatry Psychiatr Epidemiol. 2022;57(2):331-341. doi: 10.1007/s00127-021-02129-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
49.Afifi TO, Macmillan HL. Resilience following child maltreatment: a review of protective factors. Can J Psychiatry. 2011;56(5):266-272. doi: 10.1177/070674371105600505 [DOI] [PubMed] [Google Scholar]
50.Bruner MW, McLaren CD, Sutcliffe JT, et al. The effect of sport-based interventions on positive youth development: a systematic review and meta-analysis. Int Rev Sport Exerc Psychol. 2021. doi: 10.1080/1750984X.2021.1875496 [DOI] [Google Scholar]
51.Rodriguez-Ayllon M, Plaza-Florido A, Mendez-Gutierrez A, et al. The effects of a 20-week exercise program on blood-circulating biomarkers related to brain health in children with overweight or obesity: the ActiveBrains project. J Sport Health Sci. 2022;S2095-2546(22)00121-1. doi: 10.1016/j.jshs.2022.12.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
52.Frenkel D, Wilkinson K, Zhao L, et al. Scara1 deficiency impairs clearance of soluble amyloid-β by mononuclear phagocytes and accelerates Alzheimer’s-like disease progression. Nat Commun. 2013;4:2030. doi: 10.1038/ncomms3030 [DOI] [PMC free article] [PubMed] [Google Scholar]
53.Ortega FB, Mora-Gonzalez J, Cadenas-Sanchez C, et al. Effects of exercise on brain health outcomes in children with overweight/obesity: the ActiveBrains randomized controlled trial. JAMA Netw Open. 2022;5(8):e2227893. doi: 10.1001/jamanetworkopen.2022.27893 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials