Video Abstract

Video Abstract

CONTEXT:

Advances in neuroimaging techniques have resulted in an exponential increase in the number of studies investigating the effects of physical activity on brain structure and function. Authors of studies have linked physical activity and fitness with brain regions and networks integral to cognitive function and scholastic performance in children and adolescents but findings have not been synthesized.

OBJECTIVE:

To conduct a systematic review of studies in which the impact of physical activity on brain structure and function in children and adolescents is examined.

DATA SOURCES:

Six electronic databases (PubMed, PsychINFO, Scopus, Ovid Medline, SportDiscus, and Embase) were systematically searched for experimental studies published between 2002 and March 1, 2019.

STUDY SELECTION:

Two reviewers independently screened studies for inclusion according to predetermined criteria.

DATA EXTRACTION:

Two reviewers independently extracted data for key variables and synthesized findings qualitatively.

RESULTS:

Nine studies were included (task-based functional MRI [n = 4], diffusion tensor imaging [n = 3], arterial spin labeling [n = 1], and resting-state functional MRI [n = 1]) in which results for 5 distinct and 4 similar study samples aged 8.7 ± 0.6 to 10.2 ± 1.0 years and typically of relatively low socioeconomic status were reported. Effects were reported for 12 regions, including frontal lobe (n = 3), parietal lobe (n = 3), anterior cingulate cortex (n = 2), hippocampus (n = 1), and several white matter tracts and functional networks.

LIMITATIONS:

Findings need to be interpreted with caution as quantitative syntheses were not possible because of study heterogeneity.

CONCLUSIONS:

There is evidence from randomized controlled trials that participation in physical activity may modify white matter integrity and activation of regions key to cognitive processes. Additional larger hypothesis-driven studies are needed to replicate findings.

Many children and adolescents are not sufficiently active to accrue the extensive cardiovascular, metabolic, musculoskeletal, and mental health benefits of physical activity.1,2  Habitual physical activity is associated with a variety of health-related fitness traits (ie, cardiorespiratory, morphologic, muscular, motor, and metabolic),3  and emerging evidence suggests that participation in physical activity and improving physical fitness may enhance cognitive health across the life span.49 

Specifically, acute physical activity can enhance children’s attention (g = 0.43; 95% confidence interval [CI] = 0.09–0.77) and on-task behavior in the classroom (d = 0.77; 95% CI = 0.22–1.32).1012  Similarly, authors of experimental studies have demonstrated longer-term benefits of physical activity for executive functions (g = 0.24; 95% CI = 0.09–0.39),11  attention (g = 0.90; 95% CI = 0.56–1.24),11  and academic performance (g = 0.26; 95% CI = 0.02–0.49).5,11,13  Higher levels of cardiorespiratory fitness are also positively associated with young people’s academic achievement.13  Although awareness of the positive effects of physical activity on cognitive and/or academic outcomes has increased rapidly in the last 5 years, the mechanisms responsible remain relatively untested in young people.14 

Animal studies have provided initial insight into the neurobiological changes induced by physical activity. Molecular effects include epigenetic regulation of gene expression and related changes in concentrations of factors such as brain-derived neurotrophic factor (BDNF) and vascular endothelial growth factor, known to underpin brain plasticity and cellular changes such as neurogenesis, synaptogenesis, and angiogenesis.1519  There is now empirical evidence that the same molecular effects exist in humans (eg, increases in BDNF and vascular endothelial growth factor) and may be responsible for the positive effects of physical activity on cognitive health.2023 

In addition, a seminal randomized controlled trial (RCT) in older adults demonstrated that 12 months of aerobic exercise increased hippocampal volume and improved memory, with these improvements being mediated by increases in BDNF.24  Since the publication of these findings, there has been an exponential increase in the number of studies employing MRI techniques to examine associations and explore the impact of physical activity on brain structure and function in humans. Authors of many cross-sectional studies have linked physical activity with brain regions and networks integral to cognitive function and scholastic performance in children and adolescents.2528 

To date, there has been no systematic review of experimental MRI studies in which the impact of physical activity on brain structure and function in children and adolescents is investigated. A recent review of 84 studies in which the effects of physical activity on cognitive functioning and neuroimaging findings were investigated only included 5 MRI studies because the search was conducted in July 2017 and it only included RCTs.29  To provide a more in-depth and up-to-date summary of evidence of MRI studies specifically, our review included all designs of experimental studies. Given the importance of cognitive development, clarifying the effects of physical activity on brain structure and function may motivate key stakeholders to address the current physical inactivity pandemic. Therefore, our aim with this study was to conduct a systematic review of MRI studies in which the impact of physical activity on brain structure and function in school-aged children have been examined.

The conduct and reporting of this review adhere to the guidelines outlined in the Preferred Reporting Items for Systematic Reviews and Meta-Analysis statement.30  The review protocol was registered with the International Prospective Register of Systematic Reviews (CRD42017081804).

  1. Types of participants: participants were typically developing school-aged children (usually 5–18 years of age; however, children outside this age range were included if they were recruited within schools). Studies including populations with learning difficulties, cognitive deficits, and developmental disorders were excluded.

  2. Types of studies: experimental studies were eligible if the authors reported statistical analyses of changes in brain structure or function before and after a physical activity intervention.

  3. Measure of physical activity, cardiorespiratory fitness, or muscular fitness: studies with objective (eg, accelerometers and pedometers) or subjective measures of physical activity (eg, exercise session attendance and self-report questionnaires); cardiorespiratory fitness (eg, maximum oxygen consumption [VO2max] test, Progressive Aerobic Cardiovascular Endurance Run, and predictive equations); and/or muscular fitness (eg, dynamometry, standing long jump, and push up test) were eligible.

  4. Brain imaging techniques: studies that reported findings from MRI techniques (eg, functional MRI [fMRI], diffusion tensor imaging [DTI], and arterial spin labeling [ASL]) that have been used to identify structural and functional mechanisms that may explain the relationship between physical activity, cardiorespiratory fitness or muscular fitness, and cognition or academic achievement were eligible.

Six electronic databases (PubMed, PsychINFO, Scopus, Ovid Medline, SportDiscus, and Embase) were searched for studies published within the last 16 years (2002–March 1, 2019) (Supplemental Table 4). Additional searches of recently published systematic reviews in which the associations between physical activity, cardiorespiratory fitness or muscular fitness, and cognitive outcomes were examined were conducted, and the reference lists of all retrieved studies were reviewed. The search was restricted to articles published in the English language.

The study screening and selection process was performed on Covidence.31  One reviewer screened the titles and abstracts of records retrieved by the search strategy and classified these as possibly relevant or definitely irrelevant. The full-text articles of records classified as possibly relevant were retrieved and independently reviewed by 2 reviewers. Studies were classified as include or exclude. If there was disagreement between reviewers, consensus was sought through discussion. Reasons were provided for excluding any possibly relevant studies.

Both reviewers independently extracted data from included studies into a purpose-built data extraction template in excel. Data extraction included (1) sample data (including sample size, age, sex, and education); (2) study details (location, design, setting, duration, and assessment points); (3) assessment of physical activity, cardiorespiratory, and/or muscular fitness (objective, subjective, laboratory-based, or field-based); (4) neuroimaging data (MRI modality, analysis methods, regions of interest, etc); (5) data analysis (statistical methods used, confounders adjusted for, etc); and (6) study findings (quantitative and qualitative).

All studies were independently assessed by 2 reviewers and were scored as low, high, or unclear for 8 criteria according to the Cochrane collaboration risk of bias tool and scoring.32  Any disagreement concerning risk of bias assessment between the 2 reviewers was resolved through discussion. Consensus was reached on all articles included in the review.

Figure 1 displays the flow of studies through the review process. After the exclusion of duplicates, the systematic search yielded 9508 potentially relevant citations, of which 153 were retained for full-text review. There was almost-perfect interrater agreement for the full-text review (κ = 0.97).33  A total of 9 articles satisfied the inclusion criteria and were included in the review, reporting results for 5 distinct and 4 similar samples of participants ranging from 8.7 ± 0.634  to 10.2 ± 1.035  years of age and typically of relatively low socioeconomic status. The sample size ranged from 936  to 143.34  The studies were conducted in North America (8 in the United States) and Asia (1 in China). Of the included studies, 7 were RCTs and 2 were acute before and after studies. Detailed information about each included study is presented in Table 1.

FIGURE 1

Preferred Reporting Items for Systematic Reviews and Meta-Analysis flow diagram.

FIGURE 1

Preferred Reporting Items for Systematic Reviews and Meta-Analysis flow diagram.

TABLE 1

Summary of Included Studies

First Author and YearStudy and Sample Characteristics (Design, n, Age (y), % Male Sex, SES, Country)ExposureMeasure of Physical Activity and/or FitnessImaging Technique and AnalysisConfounders Adjusted for in AnalysesKey Findings
Frontal lobe: executive processes, cognition, attention, and language processing       
 Chaddock-Heyman et al 201337  RCT, 23, 8.9 ± 5.8, 71, 2.0 ± 0.9,a United States 2 h (76.8 min MVPA) aerobic and muscle- and/or bone-strengthening activities after each school day for 150 out of 170 school days. Mean (SD) attendance = 82% ± 13.3% VO2max (modified Balke) Task-based fMRI: FSL, ROI approach, motion correction via a rigid body algorithm in MCFLIRT. Primary threshold level input (z): >6.00; corrected cluster significance threshold: P < .05; familywise α level: P = .05 Baseline Intervention participants showed decreases in fMRI brain activation in the right anterior prefrontal cortex (Z = 6.2) during a flanker task 
 Davis et al 201138  RCT, 19, 9.6 ± 1.0, 58, not reported, United States Daily afterschool exercise program including running games, jump rope, and modified basketball and soccer at 166 ± 8 beats per minute (∼79% HRmax). 20–40 min/d for 14 ± 1.7 wk. Mean attendance = 85% ± 13% HR monitors and attendance Task-based fMRI: AFNI and ROI approach. Volumes were registered to a representative volume, and 6 regressors were calculated (rotational and translational head motion in 3 planes), Monte Carlo simulations threshold-cluster method familywise α at P = .05 preserved with individual voxel threshold at P = .05 and a cluster size of 40 voxels Baseline Increased prefrontal (and decreased posterior parietal) cortex activity during antisaccade performance was observed in the exercise group 
 Krafft et al 201439  RCT, 43, 9.8 ± 0.8, 35, 4.9 ± 1.1,b United States 8 mo instructor-led afterschool intervention (eg, tag and jump rope) 40 min each school day at 161 beats per minute (mean; ∼77% age-predicted HRmaxVO2peak (modified Balke) Task-based fMRI: AFNI, whole-brain approach. Volumes were registered to a representative volume and regressed for rotation in x, y, and z planes. Monte Carlo simulations threshold-cluster method: familywise α of .05 preserved with 3D cluster size of 35 (antisaccade) or 37 (flanker) voxels. Not reported Exercise led to decreased activation in several prefrontal (and parietal) regions supporting antisaccade performance, including bilateral precentral gyrus, medial frontal gyrus, paracentral lobule, and right inferior frontal gyrus. The exercise group also showed increased activation in several regions supporting flanker performance, including superior frontal gyrus and the anterior cingulate 
Parietal lobe: perception and integration of somatosensory information       
 Chen et al 201636  Acute before and after, 9, 10, 56, not reported, China 30 min cycling at 60%–69% age-predicted HRmax HR monitors Task-based fMRI: SPM8, whole-brain approach, motion corrected. Statistical threshold: P < .025; cluster size threshold = 100 voxels, equivalent to cluster-level P < .05. AlphaSim corrected Baseline Acute moderate-intensity aerobic exercise benefited performance in the n-back task, increasing brain activities of the left parietal cortex (T = 8.64), right parietal cortex (T = 6.57), left hippocampus (T = 8.23), left cerebellum (T = 7.18), and right cerebellum (T = 6.47) 
 Davis et al 201138  RCT, 19, 9.6 ± 1.0, 58, not reported, United States Daily afterschool exercise program including running games, jump rope, and modified basketball and soccer at 166 ± 8 beats per minute (∼79% HRmax) for 20–40 min/d for 14 ± 1.7 wk. Mean attendance = 85% ± 13% HR monitors and attendance Task-based fMRI: AFNI, ROI approach. Volumes were registered to a representative volume, and 6 regressors were calculated (rotational and translational head motion in 3 planes). Monte Carlo simulations threshold-cluster method familywise α at P = .05, preserved with individual voxel threshold at P = .05 and a cluster size of 40 voxels Baseline Decreased posterior parietal cortex (and increased prefrontal cortex) activity during antisaccade performance was observed in the exercise group 
 Krafft et al 201439  RCT, 9.8 ± 0.8 , 35, 4.9 ± 1.1,b United States Instructor-led afterschool intervention (eg, tag and jump rope) 40 min daily at 161 beats per minute (mean: ∼77% age-predicted HRmaxVO2peak (modified Balke) Task-based fMRI: AFNI, whole-brain approach. Volumes were registered to a representative volume and regressed for rotation in x, y, and z planes. Monte Carlo simulations threshold-cluster method: familywise α of .05 preserved with 3D cluster size of 35 (antisaccade) or 37 (flanker) voxels Not reported Exercise led to decreased activation in several parietal (and prefrontal) regions supporting antisaccade performance, including superior parietal lobule, inferior parietal lobule, paracentral lobule, postcentral gyrus, and left precuneus 
Anterior cingulate cortex: executive function       
 Chaddock-Heyman et al 201337  RCT, 23, 8.9 ± 5.8, 71, 2.0 ± 0.9,a United States 2 h (76.8 min MVPA) aerobic and muscle- and/or bone-strengthening activities after each school day for 150 out of 170 school days. Mean (SD) attendance = 82% ± 13.3% VO2max (modified Balke) Task-based fMRI: FSL, ROI approach. Motion correction via a rigid body algorithm in MCFLIRT. Primary threshold level input (z): >6.00; corrected cluster significance threshold: P < .05; familywise α level: P = .05 Baseline Intervention participants showed no changes in fMRI brain activation of the anterior cingulate cortex (z = 7.1) during a flanker task 
 Krafft et al 201439  RCT, 43, 9.8 ± 0.8, 35, 4.9 ± 1.1,b United States Instructor-led after school intervention (eg, tag and jump rope) 40 min daily at 161 beats per minute (mean: ∼77% age-predicted HRmaxVO2peak (modified Balke) Task-based fMRI: AFNI, whole-brain approach. Volumes were registered to a representative volume and regressed for rotation in x, y, & z planes. Monte Carlo simulations threshold-cluster method: familywise α of .05 preserved with 3D cluster size of 35 (antisaccade) or 37 (flanker) voxels Not reported Exercise led to decreased activation in the anterior cingulate cortex (as well as the several prefrontal and parietal regions) during antisaccade performance. The exercise group also showed increased activation in several regions supporting flanker performance, including the anterior cingulate and superior frontal gyrus 
Hippocampus: memory and spatial navigation       
 Chen et al 201636  Acute before and after, 9, 10, 56, not reported, China 30 min cycling at 60%–69% age-predicted HRmax HR monitors Task-based fMRI: SPM8, whole-brain approach, motion corrected. Statistical threshold: P < .025; cluster size threshold = 100 voxels, equivalent to cluster-level P < .05. AlphaSim corrected Baseline Acute moderate-intensity aerobic exercise benefited performance in the n-back task, increasing brain activities of the left parietal cortex (T = 8.64), right parietal cortex (T = 6.57), left hippocampus (T = 8.23), left cerebellum (T = 7.18), and right cerebellum (T = 6.47) 
Cerebellum: coordination of voluntary movement, motor learning, balance, and sequence learning       
 Chen et al 201636  Acute before and after, 9, 10, 56, not reported, China 30 min cycling at 60%–69% age-predicted HRmax HR monitors Task-based fMRI: SPM8, whole-brain approach. Statistical threshold: P < .025; cluster size threshold = 100 voxels, equivalent to cluster-level P < .05. AlphaSim corrected Baseline Acute moderate-intensity aerobic exercise benefited performance in the n-back task, increasing brain activities of the left parietal cortex (T = 8.64), right parietal cortex (T = 6.57), left hippocampus (T = 8.23), left cerebellum (T = 7.18), and right cerebellum (T = 6.47) 
Functional networks       
 Krafft et al 201440  RCT, 22, 9.5 ± 0.7, 32, 4.6 ± 1.2,b United States 8 mo instructor-led afterschool intervention (eg, tag and jump rope) 40 min each school day at 164 beats per minute (mean; ∼78% age-predicted HRmaxHR monitors Resting-state fMRI: FSL, ICA approach. Visually inspected for absolute motion >1-mm shift, components representing noise were removed, and 6 motion time courses (estimated rotation and shift in x, y, and z planes) were removed. Uncorrected voxel threshold = P < .0001. Familywise α of .05 preserved with 3D clusters of ≥169 voxels Not reported Results showed a pattern of decreased synchrony after exercise training with 3 RSNs (default mode network, cognitive control, and motor). Although the motor network showed decreased synchrony in the exercise group with the cuneus, the motor network was the only RSN to also show an opposing pattern of increased synchrony within the exercise group. 
 Pontifex et al 201835  Acute crossover, 41, 10.2 ± 1.0, 56, 0.8 ± 0.2, United States 20 min fast walk and slow jog on a treadmill at 70% age-predicted HRmax HR monitors ASL: AFNI and FSL, whole-brain and ROI approaches. Control-label perfusion weighted difference images were linearly aligned to proton-density weighted images and coregistered to subject- and session-specific T1-weighted images None (no correlation between change in CBF and age, sex, pubertal status, IQ, or change in HR or blood pressure) Findings revealed no differences in CBF after the cessation of exercise relative to the active control condition across each of the networks examined (frontoparietal, executive control, and motor) 
White matter integrity       
 Chaddock-Heyman et al 201834  RCT, 143, 8.7 ± 0.55, 49, 1.91 ± 0.78,a United States 2 h after each school day for 150 d of the 170-d school year. There was 30–35 min of sustained MVPA and 90 min of intermittent MVPA. VO2max (modified Balke), HR monitors, and accelerometers DTI: FSL, FDT and TBSS, and ROI approach. Motion and eddy current corrected; skeleton threshold at FA >0.20 Not reported (no baseline group differences for age, sex, race, IQ, SES, pubertal timing, VO2max, and BMI) PA group had increased FA and decreased RD in the genu of the corpus callosum from pre- to post-test, with no changes in axonal fiber diameter. No changes in WMI in the waitlist control group. 
 Krafft et al 201441  RCT, 18, 9.7 ± 0.7, 50, 4.6 ± 1.2,b United States 8 mo instructor-led afterschool intervention (eg, tag and jump rope) 40 min each school day at 161 beats per minute (mean; ∼77% age-predicted HRmaxHR monitors DTI: FSL and Explore DTI; ROI approach. Visual inspection for and removal of motion-distorted volumes; eddy current corrected. Thresholding was not reported Age and sex Intervention did not increase SLF WMI, but higher attendance at exercise sessions, higher intensity, and greater total dose of exercise were all associated with increased SLF WMI (increased FA and decreased RD) in a dose-response manner 
 Schaeffer et al 201442  RCT, 18, 9.7 ± 0.7, not reported, not reported, United States 8 mo of 40 min of instructor-led aerobic activities (eg, tag or jump rope) every school day. Mean (SD) attendance = 60 (30)%, HR = 161 (8) beats per minute, intensity = 6.3 (1.6) METs HR monitor, VO2peak (modified Balke) DTI: FSL and Explore DTI; ROI approach. Visual inspection for and removal of motion-distorted volumes; eddy current corrected. Thresholding was not reported Race and sex The exercise group showed significantly greater positive change in bilateral uncinate FA than the sedentary group. The exercise group also showed a greater negative change in left uncinate fasciculus RD 
First Author and YearStudy and Sample Characteristics (Design, n, Age (y), % Male Sex, SES, Country)ExposureMeasure of Physical Activity and/or FitnessImaging Technique and AnalysisConfounders Adjusted for in AnalysesKey Findings
Frontal lobe: executive processes, cognition, attention, and language processing       
 Chaddock-Heyman et al 201337  RCT, 23, 8.9 ± 5.8, 71, 2.0 ± 0.9,a United States 2 h (76.8 min MVPA) aerobic and muscle- and/or bone-strengthening activities after each school day for 150 out of 170 school days. Mean (SD) attendance = 82% ± 13.3% VO2max (modified Balke) Task-based fMRI: FSL, ROI approach, motion correction via a rigid body algorithm in MCFLIRT. Primary threshold level input (z): >6.00; corrected cluster significance threshold: P < .05; familywise α level: P = .05 Baseline Intervention participants showed decreases in fMRI brain activation in the right anterior prefrontal cortex (Z = 6.2) during a flanker task 
 Davis et al 201138  RCT, 19, 9.6 ± 1.0, 58, not reported, United States Daily afterschool exercise program including running games, jump rope, and modified basketball and soccer at 166 ± 8 beats per minute (∼79% HRmax). 20–40 min/d for 14 ± 1.7 wk. Mean attendance = 85% ± 13% HR monitors and attendance Task-based fMRI: AFNI and ROI approach. Volumes were registered to a representative volume, and 6 regressors were calculated (rotational and translational head motion in 3 planes), Monte Carlo simulations threshold-cluster method familywise α at P = .05 preserved with individual voxel threshold at P = .05 and a cluster size of 40 voxels Baseline Increased prefrontal (and decreased posterior parietal) cortex activity during antisaccade performance was observed in the exercise group 
 Krafft et al 201439  RCT, 43, 9.8 ± 0.8, 35, 4.9 ± 1.1,b United States 8 mo instructor-led afterschool intervention (eg, tag and jump rope) 40 min each school day at 161 beats per minute (mean; ∼77% age-predicted HRmaxVO2peak (modified Balke) Task-based fMRI: AFNI, whole-brain approach. Volumes were registered to a representative volume and regressed for rotation in x, y, and z planes. Monte Carlo simulations threshold-cluster method: familywise α of .05 preserved with 3D cluster size of 35 (antisaccade) or 37 (flanker) voxels. Not reported Exercise led to decreased activation in several prefrontal (and parietal) regions supporting antisaccade performance, including bilateral precentral gyrus, medial frontal gyrus, paracentral lobule, and right inferior frontal gyrus. The exercise group also showed increased activation in several regions supporting flanker performance, including superior frontal gyrus and the anterior cingulate 
Parietal lobe: perception and integration of somatosensory information       
 Chen et al 201636  Acute before and after, 9, 10, 56, not reported, China 30 min cycling at 60%–69% age-predicted HRmax HR monitors Task-based fMRI: SPM8, whole-brain approach, motion corrected. Statistical threshold: P < .025; cluster size threshold = 100 voxels, equivalent to cluster-level P < .05. AlphaSim corrected Baseline Acute moderate-intensity aerobic exercise benefited performance in the n-back task, increasing brain activities of the left parietal cortex (T = 8.64), right parietal cortex (T = 6.57), left hippocampus (T = 8.23), left cerebellum (T = 7.18), and right cerebellum (T = 6.47) 
 Davis et al 201138  RCT, 19, 9.6 ± 1.0, 58, not reported, United States Daily afterschool exercise program including running games, jump rope, and modified basketball and soccer at 166 ± 8 beats per minute (∼79% HRmax) for 20–40 min/d for 14 ± 1.7 wk. Mean attendance = 85% ± 13% HR monitors and attendance Task-based fMRI: AFNI, ROI approach. Volumes were registered to a representative volume, and 6 regressors were calculated (rotational and translational head motion in 3 planes). Monte Carlo simulations threshold-cluster method familywise α at P = .05, preserved with individual voxel threshold at P = .05 and a cluster size of 40 voxels Baseline Decreased posterior parietal cortex (and increased prefrontal cortex) activity during antisaccade performance was observed in the exercise group 
 Krafft et al 201439  RCT, 9.8 ± 0.8 , 35, 4.9 ± 1.1,b United States Instructor-led afterschool intervention (eg, tag and jump rope) 40 min daily at 161 beats per minute (mean: ∼77% age-predicted HRmaxVO2peak (modified Balke) Task-based fMRI: AFNI, whole-brain approach. Volumes were registered to a representative volume and regressed for rotation in x, y, and z planes. Monte Carlo simulations threshold-cluster method: familywise α of .05 preserved with 3D cluster size of 35 (antisaccade) or 37 (flanker) voxels Not reported Exercise led to decreased activation in several parietal (and prefrontal) regions supporting antisaccade performance, including superior parietal lobule, inferior parietal lobule, paracentral lobule, postcentral gyrus, and left precuneus 
Anterior cingulate cortex: executive function       
 Chaddock-Heyman et al 201337  RCT, 23, 8.9 ± 5.8, 71, 2.0 ± 0.9,a United States 2 h (76.8 min MVPA) aerobic and muscle- and/or bone-strengthening activities after each school day for 150 out of 170 school days. Mean (SD) attendance = 82% ± 13.3% VO2max (modified Balke) Task-based fMRI: FSL, ROI approach. Motion correction via a rigid body algorithm in MCFLIRT. Primary threshold level input (z): >6.00; corrected cluster significance threshold: P < .05; familywise α level: P = .05 Baseline Intervention participants showed no changes in fMRI brain activation of the anterior cingulate cortex (z = 7.1) during a flanker task 
 Krafft et al 201439  RCT, 43, 9.8 ± 0.8, 35, 4.9 ± 1.1,b United States Instructor-led after school intervention (eg, tag and jump rope) 40 min daily at 161 beats per minute (mean: ∼77% age-predicted HRmaxVO2peak (modified Balke) Task-based fMRI: AFNI, whole-brain approach. Volumes were registered to a representative volume and regressed for rotation in x, y, & z planes. Monte Carlo simulations threshold-cluster method: familywise α of .05 preserved with 3D cluster size of 35 (antisaccade) or 37 (flanker) voxels Not reported Exercise led to decreased activation in the anterior cingulate cortex (as well as the several prefrontal and parietal regions) during antisaccade performance. The exercise group also showed increased activation in several regions supporting flanker performance, including the anterior cingulate and superior frontal gyrus 
Hippocampus: memory and spatial navigation       
 Chen et al 201636  Acute before and after, 9, 10, 56, not reported, China 30 min cycling at 60%–69% age-predicted HRmax HR monitors Task-based fMRI: SPM8, whole-brain approach, motion corrected. Statistical threshold: P < .025; cluster size threshold = 100 voxels, equivalent to cluster-level P < .05. AlphaSim corrected Baseline Acute moderate-intensity aerobic exercise benefited performance in the n-back task, increasing brain activities of the left parietal cortex (T = 8.64), right parietal cortex (T = 6.57), left hippocampus (T = 8.23), left cerebellum (T = 7.18), and right cerebellum (T = 6.47) 
Cerebellum: coordination of voluntary movement, motor learning, balance, and sequence learning       
 Chen et al 201636  Acute before and after, 9, 10, 56, not reported, China 30 min cycling at 60%–69% age-predicted HRmax HR monitors Task-based fMRI: SPM8, whole-brain approach. Statistical threshold: P < .025; cluster size threshold = 100 voxels, equivalent to cluster-level P < .05. AlphaSim corrected Baseline Acute moderate-intensity aerobic exercise benefited performance in the n-back task, increasing brain activities of the left parietal cortex (T = 8.64), right parietal cortex (T = 6.57), left hippocampus (T = 8.23), left cerebellum (T = 7.18), and right cerebellum (T = 6.47) 
Functional networks       
 Krafft et al 201440  RCT, 22, 9.5 ± 0.7, 32, 4.6 ± 1.2,b United States 8 mo instructor-led afterschool intervention (eg, tag and jump rope) 40 min each school day at 164 beats per minute (mean; ∼78% age-predicted HRmaxHR monitors Resting-state fMRI: FSL, ICA approach. Visually inspected for absolute motion >1-mm shift, components representing noise were removed, and 6 motion time courses (estimated rotation and shift in x, y, and z planes) were removed. Uncorrected voxel threshold = P < .0001. Familywise α of .05 preserved with 3D clusters of ≥169 voxels Not reported Results showed a pattern of decreased synchrony after exercise training with 3 RSNs (default mode network, cognitive control, and motor). Although the motor network showed decreased synchrony in the exercise group with the cuneus, the motor network was the only RSN to also show an opposing pattern of increased synchrony within the exercise group. 
 Pontifex et al 201835  Acute crossover, 41, 10.2 ± 1.0, 56, 0.8 ± 0.2, United States 20 min fast walk and slow jog on a treadmill at 70% age-predicted HRmax HR monitors ASL: AFNI and FSL, whole-brain and ROI approaches. Control-label perfusion weighted difference images were linearly aligned to proton-density weighted images and coregistered to subject- and session-specific T1-weighted images None (no correlation between change in CBF and age, sex, pubertal status, IQ, or change in HR or blood pressure) Findings revealed no differences in CBF after the cessation of exercise relative to the active control condition across each of the networks examined (frontoparietal, executive control, and motor) 
White matter integrity       
 Chaddock-Heyman et al 201834  RCT, 143, 8.7 ± 0.55, 49, 1.91 ± 0.78,a United States 2 h after each school day for 150 d of the 170-d school year. There was 30–35 min of sustained MVPA and 90 min of intermittent MVPA. VO2max (modified Balke), HR monitors, and accelerometers DTI: FSL, FDT and TBSS, and ROI approach. Motion and eddy current corrected; skeleton threshold at FA >0.20 Not reported (no baseline group differences for age, sex, race, IQ, SES, pubertal timing, VO2max, and BMI) PA group had increased FA and decreased RD in the genu of the corpus callosum from pre- to post-test, with no changes in axonal fiber diameter. No changes in WMI in the waitlist control group. 
 Krafft et al 201441  RCT, 18, 9.7 ± 0.7, 50, 4.6 ± 1.2,b United States 8 mo instructor-led afterschool intervention (eg, tag and jump rope) 40 min each school day at 161 beats per minute (mean; ∼77% age-predicted HRmaxHR monitors DTI: FSL and Explore DTI; ROI approach. Visual inspection for and removal of motion-distorted volumes; eddy current corrected. Thresholding was not reported Age and sex Intervention did not increase SLF WMI, but higher attendance at exercise sessions, higher intensity, and greater total dose of exercise were all associated with increased SLF WMI (increased FA and decreased RD) in a dose-response manner 
 Schaeffer et al 201442  RCT, 18, 9.7 ± 0.7, not reported, not reported, United States 8 mo of 40 min of instructor-led aerobic activities (eg, tag or jump rope) every school day. Mean (SD) attendance = 60 (30)%, HR = 161 (8) beats per minute, intensity = 6.3 (1.6) METs HR monitor, VO2peak (modified Balke) DTI: FSL and Explore DTI; ROI approach. Visual inspection for and removal of motion-distorted volumes; eddy current corrected. Thresholding was not reported Race and sex The exercise group showed significantly greater positive change in bilateral uncinate FA than the sedentary group. The exercise group also showed a greater negative change in left uncinate fasciculus RD 

AFNI, Analysis of Functional NeuroImages; CBF, cerebral blood flow; FA, fractional anisotropy; FDT, functional MRI of the Brain’s Diffusion Toolbox; FSL, functional MRI of the Brain Software Library; HR, heart rate; HRmax, maximum heart rate; ICA, independent component analysis; MCFLIRT, Motion Correction functional MRI of the Brain’s Linear Image Registration Tool; MVPA, moderate-to-vigorous physical activity; PA, physical activity; RD, radial diffusivity; ROI, region of interest; RSN, resting-state network; SES, socioeconomic status; SLF, superior longitudinal fasciculus; SPM8, Statistical Parametric Mapping 8; TBSS, Tract-Based Spatial Statistics; VO2peak, peak oxygen consumption; WMI, white matter integrity; 3D, three-dimensional.

a

Low: <2.

b

Parental education scale (1 = grade 7 or less; 2 = grades 8–9; 3 = grades 10–11; 4 = high school graduate; 5 = partial college; 6 = college graduate; 7 = postgraduate).

Detailed information about the risk of bias for the included studies is presented in Table 2. In summary, all 9 (100%) were deemed to be at unclear risk of selection bias, with unclear description of (1) sequence generation process, (2) concealed allocation processes, and [in 5 (56%) studies] (3) subgroup selection processes. Seven (78%) studies were deemed at unclear risk of reporting bias because of lack of availability of a protocol published by means of either an article or trial registration. Six (67%) studies were deemed at high risk of attrition bias because of significant dropout with inadequate analyses. Overall, only 2 (22%) studies scored as low risk of bias for ≥3 (of the 8 criteria.34,35  There was substantial interrater agreement for the risk of bias assessment (κ = 0.61).33 

TABLE 2

Risk of Bias Assessment

StudySequence GenerationAllocation ConcealmentParticipant BlindingAssessor BlindingPersonnel BlindingSelective Outcome ReportingIncomplete Outcome DataOther Sources of Bias
Chaddock-Heyman et al34  ?a ?a ?a b 
Chaddock-Heyman et al37  ?a ?a ?a ?a ?a b 
Chen et al36  − − − − c ?d ?a 
Davis et al38  ?a ?a − c ?d e 
Krafft et al39  ?a ?a ?a ?a ?a ?d b e,f 
Krafft et al41  ?a ?a ?a ?a ?a ?d b e,f 
Krafft et al40  ?a ?a ?a ?a ?a ?d b e,f 
Pontifex et al35  ?a ?a ?a ?d b 
Schaeffer et al42  ?a ?a ?a ?a ?a ?c e,f 
StudySequence GenerationAllocation ConcealmentParticipant BlindingAssessor BlindingPersonnel BlindingSelective Outcome ReportingIncomplete Outcome DataOther Sources of Bias
Chaddock-Heyman et al34  ?a ?a ?a b 
Chaddock-Heyman et al37  ?a ?a ?a ?a ?a b 
Chen et al36  − − − − c ?d ?a 
Davis et al38  ?a ?a − c ?d e 
Krafft et al39  ?a ?a ?a ?a ?a ?d b e,f 
Krafft et al41  ?a ?a ?a ?a ?a ?d b e,f 
Krafft et al40  ?a ?a ?a ?a ?a ?d b e,f 
Pontifex et al35  ?a ?a ?a ?d b 
Schaeffer et al42  ?a ?a ?a ?a ?a ?c e,f 

+ represents low risk of bias, ? represents unclear risk of bias, and − represents high risk of bias.

a

Unclear description in article.

b

Significant dropout with inadequate analyses.

c

Authors appeared to provide intervention and control.

d

No protocol.

e

Inadequate description of subgroup selection.

f

Risk of intervention contamination.

Four different MRI modalities were used across the 9 included studies. Four (44%) studies used task-based fMRI, 3 (33%) studies used DTI, 1 (11%) study used ASL, and 1 (11%) study used resting-state fMRI. Data for 12 regions were reported across the 9 included studies: anterior cingulate cortex, cerebellum, corpus callosum, frontal lobe, hippocampus, parietal lobe, superior longitudinal fasciculus, uncinate fasciculus, cognitive control network, default mode network, executive control network, and motor network.

Authors of seven (78%) studies provided physical activity interventions and investigated effects on brain structure or function.34,3742  The duration of the interventions ranged from 3 to 9 months and generally consisted of moderate-to-vigorous physical activity (eg, 70%–80% maximum heart rate [HRmax]) either twice a week or each school day for 20 to 120 minutes. Of these, 4 studies measured cardiorespiratory fitness by means of oxygen uptake during a maximal graded treadmill test (modified Balke protocol).34,37,39,42  Authors of 2 studies investigated changes in brain function in response to acute bouts of aerobic exercise at 60% to 70% HRmax.35,36  Details of all interventions are outlined in Table 1.

Findings from each included study are presented by brain region below and Table 1, with effects further summarized in Table 3.

TABLE 3

Summary of Studies Which Have Examined the Impact of Physical Activity on Brain Structure or Brain Function

Positively Associated With PANegatively Associated With PANot Associated With PA
Task-based fMRI
 Task-positive regions Chen et al36 ,a Krafft et al39 ,b Chaddock-Heyman et al37  
 Davis et al38 ,a — — 
 Krafft et al39 ,a — — 
 Task-negative regions — Davis et al38 ,b — 
 — Krafft et al39 ,b — 
Resting-state fMRI — Krafft et al40 ,c — 
DTI Chaddock-Heyman et al34 ,d — — 
 Krafft et al41 ,d — — 
 Schaeffer et al42 ,d — — 
ASL — — Pontifex et al35  
Positively Associated With PANegatively Associated With PANot Associated With PA
Task-based fMRI
 Task-positive regions Chen et al36 ,a Krafft et al39 ,b Chaddock-Heyman et al37  
 Davis et al38 ,a — — 
 Krafft et al39 ,a — — 
 Task-negative regions — Davis et al38 ,b — 
 — Krafft et al39 ,b — 
Resting-state fMRI — Krafft et al40 ,c — 
DTI Chaddock-Heyman et al34 ,d — — 
 Krafft et al41 ,d — — 
 Schaeffer et al42 ,d — — 
ASL — — Pontifex et al35  

PA, physical activity; —, not applicable.

a

Increased activation.

b

Decreased activation.

c

Decreased synchrony of resting-state networks with regions outside those networks.

d

Increased white matter integrity.

Frontal Lobe

Authors of 3 RCTs with distinct but similarly aged samples reported results for changes in activation of the frontal lobe in response to physical activity interventions which ranged from 20 to 77 minutes each school day over 3 to 9 months. Authors of 2 of the RCTs assessed prefrontal activation during cognitive tasks (antisaccade [n = 2] and flanker [n = 2]) and found changes pre- and post- intervention but the effects were in opposite directions in both cases. Davis et al38  reported that increased bilateral prefrontal (and decreased posterior parietal) cortex activity was observed during antisaccade performance in the physical activity group, whereas Krafft et al39  reported decreased activation during antisaccade performance in several prefrontal (and parietal) regions including medial frontal gyrus, right inferior frontal gyrus, and bilateral precentral gyrus. Krafft et al39  also observed increased activation of the superior frontal gyrus of the prefrontal cortex during incongruent trials of the flanker task in the physical activity group, whereas authors of the third RCT (Chaddock-Heyman et al37 ) observed decreased activation in the right anterior prefrontal cortex during incongruent trials of the flanker task in the physical activity intervention group but no changes in the control group. Note that although both Chaddock-Heyman et al37  and Davis et al38  adjusted for baseline during their region of interest analyses, Krafft et al39  did not report if/what covariates were adjusted for and employed a whole-brain analysis approach, which could contribute to the disparate results.

Parietal Lobe

Authors of 3 studies reported results for the parietal lobe from task-based fMRI paradigms. Authors of 2 RCTs with similarly aged samples and relatively similar type, frequency, intensity, and duration of physical activity interventions found decreased parietal cortex activity during antisaccade performance after a physical activity intervention.38,39  Both studies used comparable cluster size thresholds but it should be noted that while Davis et al38  adjusted for baseline in their analyses, Krafft et al39  did not report if and what covariates were adjusted for.

Chen et al36  investigated the acute effects of a 30-minute bout of cycling (60%–69% HRmax) during task-based fMRI and reported improved n-back performance and increased activation of bilateral parietal cortices (as well as the left hippocampus and bilateral cerebellum).

Anterior Cingulate Cortex

Authors of 2 RCTs reported task-based fMRI results for the anterior cingulate cortex. Authors of 1 RCT found that participation in a physical activity intervention did not change activation of anterior cingulate cortex during neutral or incongruent conditions of a flanker task.37  The other RCT found that although there were no significant correlations between changes in cardiorespiratory fitness and brain activation during task-based fMRI,39  the physical activity intervention led to differential activation across 2 inhibition tasks, with decreased activation of the anterior cingulate cortex during an antisaccade task and increased activation of the cingulate gyrus during the incongruent condition of a flanker task.39  Comparatively, the control group showed decreased activation during the flanker task.39  Such differences across inhibition tasks highlights the complexity of brain activation during performance of tasks that tap different aspects of a similar cognitive construct.

Hippocampus

Authors of 1 acute before and after study reported enhanced performance in an n-back task and increased brain activity (task-based fMRI) of the left hippocampus in response to an acute 30-minute bout of cycling (60%–69% HRmax).36 

Cerebellum

Authors of 1 acute experimental study investigated the effects of a 30-minute bout of cycling (60%–69% HRmax) during task-based fMRI and reported improved n-back performance and increased activation of bilateral cerebellum.36 

Functional Networks

Authors of 2 experimental studies reported results for specific functional brain networks. Authors of 1 RCT used an independent component analysis approach and reported that a physical activity intervention caused decreased synchrony between the default mode network and the cognitive control network with brain regions outside of those networks during resting-state fMRI.40  There was no change in synchrony of the salience network, whereas the motor network had decreased synchrony with the left cuneus but increased synchrony with certain frontal regions.40 

Pontifex et al35  investigated the acute effects of a 20-minute bout of fast walking and/or slow jogging (70% HRmax) on cerebral blood flow in 10.2 ± 1.0 year old (n = 41) and found no differences across any of the networks examined (frontoparietal, executive control, and motor networks).

White Matter Integrity

Authors of 3 studies reported results of 2 RCTs that had examined the effects of physical activity on white matter tracts in similarly aged children using regions of interest analyses.34,41,42  One large RCT (n = 143) revealed that 2 hours of physical activity each school day for 8 months improved white matter integrity (ie, increased fractional anisotropy, which indicates the orientation of diffusion and is higher along well-defined pathways) and decreased radial diffusivity (a marker of myelin disintegration) in the genu of the corpus callosum from pretest to post-test, with no changes in estimates of axonal fiber diameter (axial diffusivity).34  There were no changes in the white matter integrity of the wait list control group, reflective of typical development. The other RCT (n = 18) also delivered an 8-month intervention consisting of a 40-minute session each school day. Authors of 1 study reported that the physical activity group showed greater increases in bilateral uncinate fasciculus fractional anisotropy and greater decreases in left uncinate fasciculus radial diffusivity compared with the control group.42  In the second report from this RCT, the physical activity intervention did not significantly increase white matter integrity in the superior longitudinal fasciculus. However, higher attendance in the exercise intervention, higher intensity, and greater total dose of exercise were all associated with increased fractional anisotropy and decreased radial diffusivity of the superior longitudinal fasciculus in a dose-response manner.41 

In this systematic review, we examined evidence of the impact of physical activity on brain structure and function in youth from MRI studies. Nine experimental studies were included in the review, of which 7 were RCTs and 2 were acute before and after studies, reporting data for 12 regions acquired with 4 MRI modalities. All 7 RCT studies (4 samples) reported significant changes in either brain structure or function after a physical activity intervention in young people.3742 

To date, the parietal cortex is the only specific region that has had >1 RCT report in which authors found an impact of physical activity on brain structure or function and for the effects to be in the same direction (ie, authors of both RCTs found decreased posterior parietal cortex activity during antisaccade performance after a physical activity intervention38,39 ). Otherwise, RCT findings for the impact of physical activity on activation during task-based fMRI were inconsistent (ie, authors of 1 study found an association and another did not) for the anterior cingulate cortex,37,39  or conflicting (ie, physical activity had an impact on activation, but authors of 1 study reported increased activation and authors of another study reported a decreased activation in the case of each task paradigm [antisaccade and incongruent condition of a flanker task]) for frontal regions.3739  It should be noted that although the sample ages and cluster size thresholds were similar in these studies, the interventions varied from 20 to 77 minutes per session over 3 to 9 months which presents considerable heterogeneity.

The desired direction of the effect of physical activity on activation will differ depending on the region and context (eg, task and rest) of interest. However, positive associations between physical activity and activation of task-positive regions during performance of task paradigms is interpreted as a greater ability to use resources in some studies,36,43,44  whereas negative associations (ie, less activation) are considered to represent a more efficient use of resources in others.37,45  There is evidence to support decreased activation of a task-positive region during task performance being reflective of a more mature and adult-like brain4648  but this should be interpreted with caution until the findings have been replicated by studies adequately powered to perform mediation analyses.49 

Authors of 2 RCTs found that physical activity caused decreased activation of the posterior parietal cortex during antisaccade task performance. Although this did not reflect a difference in antisaccade performance between the physical activity and control group in 1 study,39  authors of the other study did not report data for antisaccade task performance.38  The inferior parietal lobule, located within the posterior parietal cortex, forms part of the default mode (task-negative) network,5052  which is known to decouple from the cognitive control network during successful performance of a cognitive task.53  Therefore, these results may indicate a more refined, adult-like pattern of activation in the exercise group while maintaining equivalent levels of task performance.5456  A recent meta-analysis revealed that deactivation of the default mode network is essential for processing information so that it can later be remembered.57  This diversion of processing resources from the default mode network to brain regions involved in the task performance has previously been demonstrated in a cross-sectional pediatric physical activity study. Despite similar memory performance to their inactive peers, during encoding of later remembered versus forgotten word pairs, participants with high levels of physical activity displayed (1) robust deactivation of the default mode network, (2) strong negative coupling with the hippocampus, and (3) a more focal increase in activation of the left hippocampus only.45 

Decreased synchrony between a given network and regions outside of that network is usually an indication of a more focal, coherent, and specialized pattern of activation.58,59  Authors of 1 RCT in this review examined deactivation and activation of functional networks during resting-state fMRI and found that physical activity may be conducive of a more mature efficient brain by causing decreased synchrony of the default mode network and cognitive control network with brain regions outside of those networks during resting-state fMRI.40 

In terms of structural changes, 1 large RCT (FITKids2; n = 143) revealed that participation in physical activity can improve white matter integrity of the corpus callosum; a region important for cognitive processing.34  A second RCT investigated effects of physical activity on white matter integrity and detected significant improvements in the bilateral uncinate fasciculus (which usually matures later than many other tracts60 ).42  This was particularly evident in the left uncinate fasciculus, which is linked with auditory-verbal memory proficiency, verbal IQ, and full-scale IQ.42,61,62 

In a second study from the same RCT,41  changes in white matter integrity of the superior longitudinal fasciculus were not significantly different between the groups. However, higher attendance at exercise sessions, higher intensity, and greater total dose of exercise were positively associated with changes white matter integrity.41  Similarly, white matter integrity did not change among adults participating in a 1-year exercise intervention, but changes in fitness were positively associated with white matter integrity of prefrontal and temporal regions (which are linked by the uncinate fasciculus).63  Improvements in fitness were also associated with changes in short-term memory, but increases in white matter integrity were not associated with short‐term memory improvement. In another larger-scale study involving adults, white matter integrity in multiple tracts (including those that connect medial temporal and prefrontal cortices) mediated the relationship between fitness and spatial working memory.64  Additional support for the importance of fitness in terms of white matter integrity also exists in pediatric cross-sectional studies, which have found positive associations between fitness and fractional anisotropy in several of the same white matter tracts in children.65 

To date, no RCT has examined the impact of a physical activity intervention on volumes of brain regions in children or adolescents. This is surprising given that a recent meta-analysis on the effect of aerobic exercise on hippocampal volume in adults included 14 studies.66  This review revealed a significant effect of aerobic exercise on both left and right hippocampal volume in comparison with control conditions in healthy older adults. The effects were driven by exercise attenuating normal age-related neurodegeneration, which has been shown to precede and lead to cognitive decline and Alzheimer disease.67,68  Whether exercise can increase the volumetric growth of the hippocampus and whether these increases in volume subsequently confer benefits to cognition, memory, and/or academic performance during childhood and adolescence has not been established.

More studies in adolescents are needed because all experimental studies included in this review were conducted with children. Future researchers should also measure cardiorespiratory and muscular fitness so that (1) baseline fitness can be adjusted for in analyses and (2) changes in fitness due to physical activity interventions can be analyzed for correlations with changes in brain structure or function. There is considerable scope for different intensities, frequencies, and types of physical activity such as high-intensity interval training, resistance exercise, exergaming, and cognitively demanding physical activity to be explored.69 

Although this is the first systematic review of MRI studies in the area of pediatric physical activity, there are some limitations that should be noted. Most notably, because of the small number of RCTs and considerable heterogeneity of included studies, we were unable to conduct meta-analyses. In addition, we did not check for a file drawer effect so the risk of publication bias cannot be ruled out.

There are a number of common study limitations that should be noted. The majority of the included studies included small samples and/or relied on statistical significance analyses. The P values do not provide an indication of the size of an effect nor the importance of a result and by themselves are not a good measure of evidence regarding a model or hypothesis.70  As such, the field needs to progress to promote the reporting of effect estimates in addition to the corresponding statistics.71  Risk of bias was largely unclear across all domains and studies. Researchers are encouraged to adhere to the Consolidated Standards of Reporting Trials guidelines72  to reduce the risk of bias, particularly in terms of selection bias and reporting bias.73  Findings need to be interpreted with caution until additional RCTs can (1) replicate findings and (2) establish whether exercise-induced changes in brain structure or function mediate the cognitive and/or academic benefits of physical activity.

There is some evidence from RCTs that participation in physical activity may enhance brain structure and function in terms of white matter integrity and activation of regions key to cognitive processes, respectively. No RCT researchers have reported on the impact of physical activity on volumes of brain regions in children or adolescents.

Dr Valkenborghs conducted the search, screening, extraction, and synthesis processes in addition to drafting the manuscript; Dr Noetel screened articles, extracted data, and critically reviewed the manuscript; Drs Hillman, Nilsson, and Smith contributed to the conceptualization of the review and critically reviewed the manuscript; Dr Ortega critically reviewed the manuscript; Dr Lubans conceptualized the review and contributed to the design, synthesis, and drafting of the manuscript; and all authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.

This trial has been registered with the International Prospective Register of Systematic Reviews (https://www.crd.york.ac.uk/prospero/) (identifier CRD42017081804).

FUNDING: Supported by an Australian Research Council Future Fellowship grant (FT 140100399).

     
  • ASL

    arterial spin labeling

  •  
  • BDNF

    brain-derived neurotrophic factor

  •  
  • CI

    confidence interval

  •  
  • DTI

    diffusion tensor imaging

  •  
  • fMRI

    functional MRI

  •  
  • HRmax

    maximum heart rate

  •  
  • RCT

    randomized controlled trial

  •  
  • VO2max

    maximum oxygen consumption

1
Janssen
I
,
Leblanc
AG
.
Systematic review of the health benefits of physical activity and fitness in school-aged children and youth
.
Int J Behav Nutr Phys Act
.
2010
;
7
(
1
):
40
2
Hallal
PC
,
Andersen
LB
,
Bull
FC
, et al;
Lancet Physical Activity Series Working Group
.
Global physical activity levels: surveillance progress, pitfalls, and prospects
.
Lancet
.
2012
;
380
(
9838
):
247
257
3
Lang
JJ
,
Tomkinson
GR
,
Janssen
I
, et al
.
Making a case for cardiorespiratory fitness surveillance among children and youth
.
Exerc Sport Sci Rev
.
2018
;
46
(
2
):
66
75
4
Biddle
SJ
,
Asare
M
.
Physical activity and mental health in children and adolescents: a review of reviews
.
Br J Sports Med
.
2011
;
45
(
11
):
886
895
5
Esteban-Cornejo
I
,
Tejero-Gonzalez
CM
,
Sallis
JF
,
Veiga
OL
.
Physical activity and cognition in adolescents: a systematic review
.
J Sci Med Sport
.
2015
;
18
(
5
):
534
539
6
Donnelly
JE
,
Hillman
CH
,
Castelli
D
, et al
.
Physical activity, fitness, cognitive function, and academic achievement in children: a systematic review
.
Med Sci Sports Exerc
.
2016
;
48
(
6
):
1197
1222
7
Ruiz-Ariza
A
,
Grao-Cruces
A
,
de Loureiro
NEM
,
Martínez-López
EJ
.
Influence of physical fitness on cognitive and academic performance in adolescents: a systematic review from 2005–2015
.
Int Rev Sport Exerc Psychol
.
2017
;
10
(
1
):
108
133
8
Costigan
SA
,
Eather
N
,
Plotnikoff
RC
,
Hillman
CH
,
Lubans
DR
.
High-intensity interval training for cognitive and mental health in adolescents
.
Med Sci Sports Exerc
.
2016
;
48
(
10
):
1985
1993
9
Lubans
DR
,
Smith
JJ
,
Morgan
PJ
, et al
.
Mediators of psychological well-being in adolescent boys
.
J Adolesc Health
.
2016
;
58
(
2
):
230
236
10
Álvarez-Bueno
C
,
Pesce
C
,
Cavero-Redondo
I
, et al
.
Academic achievement and physical activity: a meta-analysis
.
Pediatrics
.
2017
;
140
(
6
):
e20171498
11
de Greeff
JW
,
Bosker
RJ
,
Oosterlaan
J
,
Visscher
C
,
Hartman
E
.
Effects of physical activity on executive functions, attention and academic performance in preadolescent children: a meta-analysis
.
J Sci Med Sport
.
2018
;
21
(
5
):
501
507
12
Daly-Smith
AJ
,
Zwolinsky
S
,
McKenna
J
, et al
.
Systematic review of acute physically active learning and classroom movement breaks on children’s physical activity, cognition, academic performance and classroom behaviour: understanding critical design features
.
BMJ Open Sport Exerc Med
.
2018
;
4
(
1
):
e000341
13
Marques
A
,
Santos
DA
,
Hillman
CH
,
Sardinha
LB
.
How does academic achievement relate to cardiorespiratory fitness, self-reported physical activity and objectively reported physical activity: a systematic review in children and adolescents aged 6-18 years
.
Br J Sports Med
.
2018
;
52
(
16
):
1039
14
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
15
Fernandes
J
,
Arida
RM
,
Gomez-Pinilla
F
.
Physical exercise as an epigenetic modulator of brain plasticity and cognition
.
Neurosci Biobehav Rev
.
2017
;
80
:
443
456
16
Cooper
C
,
Moon
HY
,
van Praag
H
.
On the run for hippocampal plasticity
.
Cold Spring Harb Perspect Med
.
2018
;
8
(
4
):
a029736
17
Lista
I
,
Sorrentino
G
.
Biological mechanisms of physical activity in preventing cognitive decline
.
Cell Mol Neurobiol
.
2010
;
30
(
4
):
493
503
18
Vaynman
S
,
Ying
Z
,
Gomez-Pinilla
F
.
Hippocampal BDNF mediates the efficacy of exercise on synaptic plasticity and cognition
.
Eur J Neurosci
.
2004
;
20
(
10
):
2580
2590
19
Rich
B
,
Scadeng
M
,
Yamaguchi
M
,
Wagner
PD
,
Breen
EC
.
Skeletal myofiber vascular endothelial growth factor is required for the exercise training-induced increase in dentate gyrus neuronal precursor cells
.
J Physiol
.
2017
;
595
(
17
):
5931
5943
20
Hashimoto
T
,
Tsukamoto
H
,
Takenaka
S
, et al
.
Maintained exercise-enhanced brain executive function related to cerebral lactate metabolism in men
.
FASEB J
.
2018
;
32
(
3
):
1417
1427
21
Nascimento
CM
,
Pereira
JR
,
de Andrade
LP
, et al
.
Physical exercise in MCI elderly promotes reduction of pro-inflammatory cytokines and improvements on cognition and BDNF peripheral levels
.
Curr Alzheimer Res
.
2014
;
11
(
8
):
799
805
22
Leckie
RL
,
Oberlin
LE
,
Voss
MW
, et al
.
BDNF mediates improvements in executive function following a 1-year exercise intervention
.
Front Hum Neurosci
.
2014
;
8
:
985
23
Voss
MW
,
Erickson
KI
,
Prakash
RS
, et al
.
Neurobiological markers of exercise-related brain plasticity in older adults
.
Brain Behav Immun
.
2013
;
28
:
90
99
24
Erickson
KI
,
Voss
MW
,
Prakash
RS
, et al
.
Exercise training increases size of hippocampus and improves memory
.
Proc Natl Acad Sci USA
.
2011
;
108
(
7
):
3017
3022
25
Chaddock-Heyman
L
,
Weng
TB
,
Kienzler
C
, et al
.
Scholastic performance and functional connectivity of brain networks in children
.
PLoS One
.
2018
;
13
(
1
):
e0190073
26
Talukdar
T
,
Nikolaidis
A
,
Zwilling
CE
, et al
.
Aerobic fitness explains individual differences in the functional brain connectome of healthy young adults
.
Cereb Cortex
.
2018
;
28
(
10
):
3600
3609
27
Chaddock
L
,
Erickson
KI
,
Prakash
RS
, et al
.
A neuroimaging investigation of the association between aerobic fitness, hippocampal volume, and memory performance in preadolescent children
.
Brain Res
.
2010
;
1358
:
172
183
28
Bunketorp Käll
L
,
Malmgren
H
,
Olsson
E
,
Lindén
T
,
Nilsson
M
.
Effects of a curricular physical activity intervention on children’s school performance, wellness, and brain development
.
J Sch Health
.
2015
;
85
(
10
):
704
713
29
Gunnell
KE
,
Poitras
VJ
,
LeBlanc
A
, et al
.
Physical activity and brain structure, brain function, and cognition in children and youth: a systematic review of randomized controlled trials
.
Ment Health Phys Act
.
2019
;
16
:
105
127
30
Liberati
A
,
Altman
DG
,
Tetzlaff
J
, et al
.
The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration
.
Ann Intern Med
.
2009
;
151
(
4
):
W65
W94
31
Covidence. Veritas Health Innovation, Melbourne, Australia. Available at: www.covidence.org. Accessed August 16, 2019
32
Higgins
JP
,
Altman
DG
,
Gøtzsche
PC
, et al;
Cochrane Bias Methods Group
;
Cochrane Statistical Methods Group
.
The Cochrane Collaboration’s tool for assessing risk of bias in randomised trials
.
BMJ
.
2011
;
343
:
d5928
33
Landis
JR
,
Koch
GG
.
The measurement of observer agreement for categorical data
.
Biometrics
.
1977
;
33
(
1
):
159
174
34
Chaddock-Heyman
L
,
Erickson
KI
,
Kienzler
C
, et al
.
Physical activity increases white matter microstructure in children
.
Front Neurosci
.
2018
;
12
(
950
):
950
35
Pontifex
MB
,
Gwizdala
KL
,
Weng
TB
,
Zhu
DC
,
Voss
MW
.
Cerebral blood flow is not modulated following acute aerobic exercise in preadolescent children
.
Int J Psychophysiol
.
2018
;
134
:
44
51
36
Chen
AG
,
Zhu
LN
,
Yan
J
,
Yin
HC
.
Neural basis of working memory enhancement after acute aerobic exercise: fMRI study of preadolescent children
.
Front Psychol
.
2016
;
7
:
1804
37
Chaddock-Heyman
L
,
Erickson
KI
,
Voss
MW
, et al
.
The effects of physical activity on functional MRI activation associated with cognitive control in children: a randomized controlled intervention
.
Front Hum Neurosci
.
2013
;
7
:
72
38
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
98
39
Krafft
CE
,
Schwarz
NF
,
Chi
L
, et al
.
An 8-month randomized controlled exercise trial alters brain activation during cognitive tasks in overweight children
.
Obesity (Silver Spring)
.
2014
;
22
(
1
):
232
242
40
Krafft
CE
,
Pierce
JE
,
Schwarz
NF
, et al
.
An eight month randomized controlled exercise intervention alters resting state synchrony in overweight children
.
Neuroscience
.
2014
;
256
:
445
455
41
Krafft
CE
,
Schaeffer
DJ
,
Schwarz
NF
, et al
.
Improved frontoparietal white matter integrity in overweight children is associated with attendance at an after-school exercise program
.
Dev Neurosci
.
2014
;
36
(
1
):
1
9
42
Schaeffer
DJ
,
Krafft
CE
,
Schwarz
NF
, et al
.
An 8-month exercise intervention alters frontotemporal white matter integrity in overweight children
.
Psychophysiology
.
2014
;
51
(
8
):
728
733
43
Voss
MW
,
Chaddock
L
,
Kim
JS
, et al
.
Aerobic fitness is associated with greater efficiency of the network underlying cognitive control in preadolescent children
.
Neuroscience
.
2011
;
199
:
166
176
44
Mehta
RK
,
Shortz
AE
,
Benden
ME
.
Standing up for learning: a pilot investigation on the neurocognitive benefits of stand-biased school desks
.
Int J Environ Res Public Health
.
2015
;
13
(
1
):
ijerph13010059
45
Herting
MM
,
Nagel
BJ
.
Differences in brain activity during a verbal associative memory encoding task in high- and low-fit adolescents
.
J Cogn Neurosci
.
2013
;
25
(
4
):
595
612
46
Casey
BJ
,
Trainor
RJ
,
Orendi
JL
, et al
.
A developmental functional MRI study of prefrontal activation during performance of a go-no-go task
.
J Cogn Neurosci
.
1997
;
9
(
6
):
835
847
47
Scherf
KS
,
Sweeney
JA
,
Luna
B
.
Brain basis of developmental change in visuospatial working memory
.
J Cogn Neurosci
.
2006
;
18
(
7
):
1045
1058
48
Squire
LR
,
Ojemann
JG
,
Miezin
FM
, et al
.
Activation of the hippocampus in normal humans: a functional anatomical study of memory
.
Proc Natl Acad Sci USA
.
1992
;
89
(
5
):
1837
1841
49
Stillman
CM
,
Cohen
J
,
Lehman
ME
,
Erickson
KI
.
Mediators of physical activity on neurocognitive function: a review at multiple levels of analysis
.
Front Hum Neurosci
.
2016
;
10
:
626
50
Cabeza
R
,
Nyberg
L
.
Imaging cognition II: an empirical review of 275 PET and fMRI studies
.
J Cogn Neurosci
.
2000
;
12
(
1
):
1
47
51
McKiernan
KA
,
Kaufman
JN
,
Kucera-Thompson
J
,
Binder
JR
.
A parametric manipulation of factors affecting task-induced deactivation in functional neuroimaging
.
J Cogn Neurosci
.
2003
;
15
(
3
):
394
408
52
Shulman
GL
,
Fiez
JA
,
Corbetta
M
, et al
.
Common blood flow changes across visual tasks: II. Decreases in cerebral cortex
.
J Cogn Neurosci
.
1997
;
9
(
5
):
648
663
53
Putcha
D
,
Ross
RS
,
Cronin-Golomb
A
,
Janes
AC
,
Stern
CE
.
Salience and default mode network coupling predicts cognition in aging and Parkinson’s disease
.
J Int Neuropsychol Soc
.
2016
;
22
(
2
):
205
215
54
Domagalik
A
,
Beldzik
E
,
Fafrowicz
M
,
Oginska
H
,
Marek
T
.
Neural networks related to pro-saccades and anti-saccades revealed by independent component analysis
.
Neuroimage
.
2012
;
62
(
3
):
1325
1333
55
Beaty
RE
,
Benedek
M
,
Kaufman
SB
,
Silvia
PJ
.
Default and executive network coupling supports creative idea production
.
Sci Rep
.
2015
;
5
:
10964
56
Raichle
ME
,
MacLeod
AM
,
Snyder
AZ
, et al
.
A default mode of brain function
.
Proc Natl Acad Sci USA
.
2001
;
98
(
2
):
676
682
57
Kim
H
.
Neural activity that predicts subsequent memory and forgetting: a meta-analysis of 74 fMRI studies
.
Neuroimage
.
2011
;
54
(
3
):
2446
2461
58
Luna
B
,
Padmanabhan
A
,
O’Hearn
K
.
What has fMRI told us about the development of cognitive control through adolescence?
Brain Cogn
.
2010
;
72
(
1
):
101
113
59
Fox
MD
,
Snyder
AZ
,
Vincent
JL
, et al
.
The human brain is intrinsically organized into dynamic, anticorrelated functional networks
.
Proc Natl Acad Sci USA
.
2005
;
102
(
27
):
9673
9678
60
Lebel
C
,
Walker
L
,
Leemans
A
,
Phillips
L
,
Beaulieu
C
.
Microstructural maturation of the human brain from childhood to adulthood
.
Neuroimage
.
2008
;
40
(
3
):
1044
1055
61
Mabbott
DJ
,
Rovet
J
,
Noseworthy
MD
,
Smith
ML
,
Rockel
C
.
The relations between white matter and declarative memory in older children and adolescents
.
Brain Res
.
2009
;
1294
:
80
90
62
Constable
RT
,
Ment
LR
,
Vohr
BR
, et al
.
Prematurely born children demonstrate white matter microstructural differences at 12 years of age, relative to term control subjects: an investigation of group and gender effects
.
Pediatrics
.
2008
;
121
(
2
):
306
316
63
Voss
MW
,
Heo
S
,
Prakash
RS
, et al
.
The influence of aerobic fitness on cerebral white matter integrity and cognitive function in older adults: results of a one-year exercise intervention
.
Hum Brain Mapp
.
2013
;
34
(
11
):
2972
2985
64
Oberlin
LE
,
Verstynen
TD
,
Burzynska
AZ
, et al
.
White matter microstructure mediates the relationship between cardiorespiratory fitness and spatial working memory in older adults
.
Neuroimage
.
2016
;
131
:
91
101
65
Chaddock-Heyman
L
,
Erickson
KI
,
Holtrop
JL
, et al
.
Aerobic fitness is associated with greater white matter integrity in children
.
Front Hum Neurosci
.
2014
;
8
:
584
66
Firth
J
,
Stubbs
B
,
Vancampfort
D
, et al
.
Effect of aerobic exercise on hippocampal volume in humans: a systematic review and meta-analysis
.
Neuroimage
.
2018
;
166
:
230
238
67
Raz
N
,
Lindenberger
U
,
Rodrigue
KM
, et al
.
Regional brain changes in aging healthy adults: general trends, individual differences and modifiers
.
Cereb Cortex
.
2005
;
15
(
11
):
1676
1689
68
Jack
CR
 Jr
,
Wiste
HJ
,
Vemuri
P
, et al;
Alzheimer’s Disease Neuroimaging Initiative
.
Brain beta-amyloid measures and magnetic resonance imaging atrophy both predict time-to-progression from mild cognitive impairment to Alzheimer’s disease
.
Brain
.
2010
;
133
(
11
):
3336
3348
69
Schmidt
M
,
Jäger
K
,
Egger
F
,
Roebers
CM
,
Conzelmann
A
.
Cognitively engaging chronic physical activity, but not aerobic exercise, affects executive functions in primary school children: a group-randomized controlled trial
.
J Sport Exerc Psychol
.
2015
;
37
(
6
):
575
591
70
Wasserstein
RL
,
Lazar
NA
.
The ASA’s statement on p-values: context, process, and purpose
.
Am Stat
.
2016
;
70
(
2
):
129
133
71
Chen
G
,
Taylor
PA
,
Cox
RW
.
Is the statistic value all we should care about in neuroimaging?
Neuroimage
.
2017
;
147
:
952
959
72
Schulz
KF
,
Altman
DG
,
Moher
D
;
CONSORT Group
.
CONSORT 2010 statement: updated guidelines for reporting parallel group randomised trials
.
BMJ
.
2010
;
340
:
c332
73
Poldrack
RA
,
Baker
CI
,
Durnez
J
, et al
.
Scanning the horizon: towards transparent and reproducible neuroimaging research
.
Nat Rev Neurosci
.
2017
;
18
(
2
):
115
126

Competing Interests

POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.

FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.

Supplementary data