Which indices of cardiorespiratory fitness are more strongly associated with brain health in children with overweight/obesity?

To compare the strength of associations between different indices of cardiorespiratory fitness (CRF) and brain health outcomes in children with overweight/obesity.


| INTRODUCTION
Nearly 30% of children in European countries are living with overweight/obesity, and the prevalence may be increasing. 1 Childhood overweight/obesity is associated with impaired brain health. 2Some evidence also suggests that higher levels of cardiorespiratory fitness (CRF) are associated with improved executive functions, academic performance, and enhanced brain characteristics, such as improved functional connectivity and increased hippocampal volume (hereafter referred to as brain health outcomes) in children and adolescents. 3,4However, the variety of methodologies used to assess CRF 3,4 may have clouded our understanding of the importance of CRF for young people's brain health outcomes.
Previous research has identified positive associations between CRF assessed using the 20-metre shuttle run (20mSRT) or treadmill exercise tests and brain health outcomes in youth. 5][8][9][10][11] Moreover, CRF assessed by a 1.6km run, but not CRF assessed by physical work capacity at a heart rate of 170, was positively associated with academic performance in a random sample of apparently healthy children. 7This issue is further complicated with the different equations used to calculate VȮ 2peak from indirect submaximal and maximal exercise testing. 12For example, there are at least 15 different equations used to estimate VȮ 2peak using results from the 20mSRT. 13The use of different estimation equations may increase the uncertainty in the associations between CRF and brain health outcomes. 10st studies that have explored associations between directly measured VȮ 2peak and brain health outcomes have used VȮ 2peak normalized for body mass (BM). 14owever, the ratio standard approach has been criticized in the literature because it is often invalid in removing the effect of body size on VȮ 2peak 15 leading to underestimated VȮ 2peak in children with overweight/obesity. 15Therefore, normalizing the indicators of CRF using allometrically modeled lean body mass (LBM) has been recommended. 16ome studies suggest no association between VȮ 2peak normalized for LBM and brain health outcomes, 6,17 yet others have observed such a link. 18To our knowledge, the effects of the scaling approach on the associations between directly measured VȮ 2peak and brain health outcomes have yet to be comprehensively investigated in children with increased adiposity.Compared to maximal exercise, submaximal exercise testing is better tolerated, safer, and more appropriate for children with increased adiposity. 19Therefore, submaximal testing may be a more feasible approach for investigating the link between CRF and brain health outcomes in this population, but submaximal indices of CRF have been almost completely omitted in previous studies. 3,18VȮ 2 at the ventilatory threshold (VT) is widely used submaximal indicator of CRF 20 and could have a different association with brain health outcomes than VȮ 2peak .However, to the best of our knowledge, only one study has investigated the associations between VȮ 2 at the VT and brain health outcomes in youth with normal weight, showing a positive association between VȮ 2 at the VT and brain health outcomes. 18hile we have previously observed that laps completed in the 20-metre shuttle run test and estimated

K E Y W O R D S
brain, child, cognition, pediatric obesity, physical fitness VȮ 2peak are positively associated with brain health outcomes in children with overweight/obesity, [8][9][10][11] no previous study has comprehensively investigated the magnitude of the associations between several indices of CRF and a broad set of brain health outcomes.The present study contributes to the existing evidence by comparing the associations of different CRF indices with brain health outcomes, specifically focusing on (1) laboratory versus field-based; (2) measured versus estimated VȮ 2peak ; (3) different equations to estimate VȮ 2peak ; (4) different scaling approaches; (5) VȮ 2peak versus peak performance; and (6) maximal versus submaximal indices.Therefore, the aim of our study is to compare the strength of associations between different indices of CRF and brain health outcomes, including intelligence, executive function, academic performance, and total gray matter and hippocampal volumes in children with overweight/obesity.

| Study design and population
We used baseline data from the ActiveBrains trial (Clini calTr ial.gov ID: NCT02295072), 21 conducted in children aged 8-11 years with overweight/obesity.The recruitment occurred mainly at the pediatric units of the two main hospitals in Granada, Spain, from November 1, 2014 to June 30, 2016.The assessments in the study were carried out in three waves and over 5-6 days for different outcome measures in the following order: (1) brain health outcomes, (2) field-based physical fitness testing, (3) cardiometabolic risk factors, (4) maximal incremental treadmill test, (5) laboratory-based strength testing, questionnaires, and body composition (note that in the first wave, consisting 20% of the whole sample, body composition was assessed at the same day than the cardiometabolic risk factors).Children with any medical condition that would affect the results of the evaluations or that limit the normal capacity to do exercise were excluded.A total of 100 children (40 girls) had complete data and were included in the analyses of the present study.For the current analyses, we estimated that 97 observations were needed to observe the correlation of 0.25 at the power of 0.80 when statistical significance level was set at p < 0.05.The parents or legal guardians of the children provided written informed consent to participate in the trial.The ActiveBrains project was approved by the ethics committee of the University of Granada (Reference: 848, February 2014).
2.2 | Assessment of indices of cardiorespiratory fitness 2.2.1 | Maximal indices obtained from the incremental treadmill exercise test VȮ 2peak and a treadmill time were assessed during a maximal incremental treadmill test (h/p/cosmos sports and medical gmbh, Nussdorf-Traunstein, Germany) at the Andalusian Centre of Sports Medicine.Respiratory gas exchange was analyzed using a calibrated gas analyzer (General Electric Corp), and the breath-by-breath data were averaged over 10 s.Participants walked on a treadmill at a constant speed (4.8 km/h) with a 6% slope with grade increments of 1% every minute until volitional exhaustion.We defined a maximal effort in the incremental treadmill exercise test as meeting three out of four following criteria: achieving >85% of aged-predicted maximal heart rate, a respiratory exchange ratio of ≥1.0, volitional fatigue (i.e., >8 points in the OMNI scale), and a plateau in the VȮ 2 during the last two exercise work rates (<2.0 mL × kg BM −1 × min −1 ). 21Heart rate was measured an electrocardiogram.Before the treadmill exercise test, the OMNI scale was explained to children to ensure that they understood the meaning of each category of the scale.However, because of uncertainty about the secondary indicators of maximal effort in children, 22 we performed the analysis with the complete sample that performed the incremental treadmill exercise test and provided VȮ 2peak values.We also ran sensitivity analyses in the subsample of children that met the criteria for maximal effort.
VȮ 2peak (L/min) was ratio scaled for BM (VȮ 2peak mL × kg BM −1 × min −1 ) and allometrically modeled BM (VȮ 2peak mL × kg body mass −b × min −1 ) and LBM (VȮ 2peak mL × kg lean mass −b × min −1 ).Allometric scaling of VȮ 2peak was performed by the log-linear regression model 23 with BM or LBM as an independent variable and VȮ 2peak as a dependent variable.VȮ 2peak , BM and LBM were log-transformed, and least squares regression with the equation ln (VȮ 2peak ) = lnY/b ln(X) was used to obtain the scaling exponent b.The scaling exponent b for BM was 0.70 (95% confidence interval [CI] = 0.60-0.81)and for LBM 0.87 (95% CI = 0.77-0.98).These power function ratios removed the associations of VȮ 2peak with BM (r = −0.023,95% CI = −0.219 to 0.174, p = 0.810) and LBM (r = −0.016,95% CI = −0.212 to 0.181, p = 0.872), suggesting the validity of scaling CRF for body size.To test if the slope of the association of BM or LBM with VȮ 2peak was similar in boys and girls, we added the interaction term to the model.The interaction of sex with BM or LBM to VȮ 2peak was not statistically significant (p > 0.122).
2.2.2 | Maximal indices obtained from the 20-metre shuttle run test Twenty-metre SRT was performed at the Sport and Health University Research Institute (iMUDS), University of Granada, and supervised by experienced researchers.Participants were required to run between two lines 20-m apart while keeping pace with a prerecorded audio.The participants performed the test individually.One researcher ran with the participant to help keep the pace and reach maximal effort.The initial speed was 8.5 km/h, which was increased by 0.5 km/h each minute (1 min = approximately 1 stage).The CRF was recorded as completed laps and the speed at the final stage.We also estimated VȮ 2peak using the equations by Léger et al., 24 Mahar et al., 25 and Matsuzaka et al. 26 to assess whether using different equations to estimate VȮ 2peak can be used interchangeably to investigate the associations between estimated CRF and brain health outcomes.The equation by Léger et al. 24 is the most widely used equation to estimate VȮ 2peak in youth, allowing comparisons between previous studies.Furthermore, the equations by Mahar et al. 25 and Matsuzaka et al. 26 have been suggested to provide better prediction accuracy of VȮ 2peak than other equations. 13,252.2.3 | Submaximal indices obtained from the incremental treadmill exercise test VȮ 2 at VT was determined using the data acquired during the maximal incremental treadmill test by two sports physicians.The VT was defined as a point where the increase in the ventilatory equivalent for VȮ 2 occurs without an increment in the ventilatory equivalent for carbon dioxide production.The threshold was confirmed by inspecting the nonlinear increase in ventilation relative to oxygen uptake.VȮ 2 at VT was normalized for BM and allometrically modeled LBM using the VȮ 2peak approach. 18

| Assessment of brain health outcomes
Brain health outcomes were assessed at the Mind, Brain, and Behavior Research Centre, at the University of Granada, by trained researchers.Total intelligence was assessed using the Spanish version of Kaufman Brief Intelligence test measuring verbal and nonverbal intelligence. 21Normal scores from verbal and nonverbal intelligence subtests were used to calculate a composite intelligence score.
Executive functions including cognitive flexibility, inhibition, and working memory were assessed using the Design Fluency Test and the Trail Making Test, a modified version of the Stroop Color-Word Test (paper-pencil version), and a modified version of the Delayed Non-Matchto-Sample computerized task, respectively.The executive function composite z score was calculated as the renormalized mean of the z scores for cognitive flexibility, inhibition, and working memory. 8cademic performance was assessed using the Spanish version of the Woodcock-Johnson III Tests of Achievement.Total academic performance was defined as an overall performance based on reading, mathematics, and writing. 11otal gray matter volume (cm 3 ) and hippocampal (mm 3 ) volume were assessed by structural magnetic resonance imaging (Siemens Trio de 3 T, Magnetom Trio, Siemens Medical Systems, Erlangen, Germany).All images were collected on a 3.0 Tesla Siemens Magnetom Tim Trio scanner (Siemens Medical Solutions, Erlangen, Germany) with a 32-channel head coil.High-resolution T1-weighted images were acquired using a 3D MPRAGE (magnetization-prepared rapid gradient echo) protocol. 9Acquisition parameters were as follows: repetition time (TR) = 2300 ms, echo time (TE) = 3.1 ms, inversion time (TI) = 900 ms, flip angle = 9°, field of view (FOV) = 256 × 256, acquisition matrix = 320 × 320, 208 slices, resolution = 0.8 × 0.8 × 0.8 mm, and scan duration of 6 min and 34 s.
The MRI images were analyzed with FreeSurfer software version 5.3.0 (http:// surfer.nmr.mgh.harva rd.edu) and FMRIB's Software Library (FSL) version 5.0.7.(FMRIB analysis group, Oxford, UK).We used the standard processing pipeline known as recon-all that has been previously described and well-validated to assess total and gray matter volume [27][28][29] and a semiautomated model-based subcortical segmentation tool which uses the Bayesian framework from shape and appearance models obtained from manually segmented images of hippocampal volumes described in detail previously. 30Before preprocessing, we visually checked each individual image for acquisition artifacts and four children were excluded due to motion noise.In addition, outputs were visually inspected by two assessors and when an additional opinion was needed, another assessor inspected the outputs.
2. 4 | Assessment of body size and composition BM (kg) and height (cm) were measured using an electronic scale (SECA861, Hamburg, Germany) and a precision stadiometer (SECA225, Hamburg, Germany), respectively.Both measurements were performed twice and averaged.Dual energy X-ray absorptiometry (DXA) was used to measure whole body fat mass (kg), body fat percentage (BF%), and LBM (kg).The Norland XR-46 (software version 3.9.6,Medical System, Inc., Fort Atkinson, Wisconsin) scanner was used in the first wave (16 participants) while the Hologic Discovery Wi (software version APEX 4.0.2,Hologic Series Discovery QDR, Bedford, Massachusetts) was used in the second and third wave (84 participants).Subsequent analyses were completed by the same researcher following recommendations from the International Society of Clinical Densitometry. 31These analyses were performed separately for each DXA device to eliminate the potential error associated with using two different DXA devices.

| Other assessments
Somatic maturity status in terms of time to peak height velocity was calculated using the equations by Moore et al. 32 The participants were classified as pre-(<−1 years), circa (−1 to 1 years), and post (>1 years) peak height velocity.
Parental education was reported as: no elementary school, elementary school, middle school, high school, and university completed.Parents responses were combined into a trichotomous variable: none, one of the parents, or both had a university degree.

| Statistical analyses
Statistical analyses were performed using the SPSS statistical software, version 27.0 (IBM corp.Armonk, NY, USA).All continuous variables were checked for normality by observing histograms and using the Kolmogorov-Smirnov test.The associations between the indices of CRF and brain health outcomes were investigated using linear regression analyses adjusted for sex, somatic maturity status, and parental education.These data were further adjusted for BF%.We further investigated the modifying effects of sex and BF% on the associations between indices of CRF and brain health outcomes including sex × CRF or BF% × CRF interaction term to the model.The data were reported using standardized regression coefficients and 95% confidence intervals.We considered standardized regression coefficients between 0.10 and 0.29, between 0.30 and 0.49 are medium, and ≥0.50 to describe small, medium, and large effect sizes, respectively. 33

| Characteristics of participants
Participants' characteristics are reported in Table 1.A total of 66 of 100 (%) children met the criteria for maximal effort in the treadmill exercise test.Specifically, 95%, 29%, 88%, and 67% of children met the criteria for heart rate, respiratory exchange ratio, volitional fatigue, and plateau in the VȮ 2 , respectively.Children who did not meet the criteria for maximal effort did not differ in absolute VȮ 2peak , VȮ 2peak normalized for kg LBM, treadmill time, absolute or normalized VȮ 2 at VT, or in maximal OMNI score from those who met the criteria (p > 0.115).However, those who did not meet the criteria for maximal effort had lower peak respiratory exchange ratio (mean difference −0.05, 95% CI = −0.07 to −0.03, p < 0.001), peak heart rate (mean difference − 6.9, 95% CI = −11.7 to −2.1, p = 0.006), and higher VȮ 2peak normalized for kg BM −1 (mean difference 3.0, 95% CI = 1.1-4.9,p = 0.002) than those who met the criteria.

| Associations of indices of cardiorespiratory fitness with brain health outcomes
The associations between indices of CRF and brain health outcomes are presented in Figure 1.VȮ 2peak normalized for kg BM −1 or kg BM −0.70 and a longer treadmill time were positively associated with total gray matter volume.Brain health outcomes were not statistically significantly associated with VȮ 2peak or VȮ 2 at VT measured during the incremental treadmill exercise test.
Laps completed in the 20mSRT were positively associated with executive functions and academic performance.Higher speed at the final stage of the 20mSRT was associated with better executive functions.Estimated VȮ 2peak/Léger et al. was positively associated with intelligence, executive functions, academic performance, and gray matter volume.Higher estimated VȮ 2peak/Mahar et al. and VȮ 2peak/Matsuzaka et al. (speed) was positively associated with executive functions.All statistically significant associations were small in magnitude (standardized regression coefficient ranging from 0.183 to 0.256).
Most of the abovementioned associations between the indices of CRF with brain health outcomes remained materially unchanged after further adjustment for BF% (Table 2).However, the association of VȮ 2peak normalized for kg BM −1 or kg BM −0.70 , a treadmill time, and estimated VȮ 2peak/Léger et al. with total gray matter volume were no longer statistically significant after adjustment for BF%.

| Sex and body fat percentage as moderators of the associations between indices of cardiorespiratory fitness and brain health outcomes
In girls, completed laps and final speed on the 20mSRT, and estimated VȮ 2peak/Leger et al. were positively associated A B L E 1 Characteristics of participants.

All Girls Boys
Age (years) a 10.1 (9.2-11.0)9.9 (8.9-10.5)10. with academic performance with medium to large effect sizes S1).In boys, these associations were cally nonsignificant with small effect sizes (Table S1).Laps in the 20mSRT were directly associated with executive functions with medium effect sizes in children with lower BF% (below median) but the associations in children with higher BF% (at or above median) were weak and statistically nonsignificant (Table S2).Laps and speed at the final stage in the 20mSRT had positive association with medium effect size with academic performance in children with higher BF% but the association in children with lower BF% was weak and statistically nonsignificant.VȮ 2peak/Leger et al. was positively associated with gray matter volume with medium effect size in children with higher BF% but not in children with lower BF%.

| Sensitivity analyses
The associations between the indices of CRF and brain health outcomes remained materially unchanged after excluding children who did not reach three of four criteria for maximal effort during the incremental treadmill exercise test from the analyses (Table S3).

| DISCUSSION
Our main findings were that (1) directly measured VȮ 2peak and VȮ 2 at VT were not associated with behavioral brain health outcomes, (2) VȮ 2peak estimated from 20mSRT performance using the equation by Leger et al. in 1988 24 had stronger and more consistent associations with brain health outcomes than other indices of CRF, and (3) five out of six indices derived from the 20mSRT were positively associated with executive functions.Collectively, our findings suggest that while CRF is positively associated with brain health in children with overweight/obesity, not all measures of CRF are associated with brain health.Furthermore, the effect sizes for all associations were considered small.
Consistent with some previous studies in children,  , 24 Mahar et al., 25 and Matsuzaka et al., 26 respectively.
malleability of executive functions and their sensitivity to changes in CRF in childhood may explain why we observed the most consistent between CRF and executive functions.Nevertheless, supporting the evidence from one previous study, 7 we observed that peak performance in the field-based running tests is more strongly associated with behavioral brain health outcomes than laboratory-measured indices of CRF.Different determinants of peak performance and measured VȮ 2peak may explain our findings.Maximal stroke volume and cardiac output are the strongest determinants of VȮ 2peak , whereas a combination of VȮ 2peak , body composition, agility, motivation, and self-regulation determining 20mSRT performance may contribute to its stronger associations with brain health outcomes. 37However, VȮ 2peak normalized for body mass has been positively associated with executive functions and academic performance in children with overweight and low levels of CRF (mean VȮ 2peak = 27 mL × kg BM −1 × min −1 ). 14While the reason for these contrasting findings is unclear, a higher directly measured VȮ 2peak may contribute to brain health, particularly in children with very low CRF.Furthermore, the larger sample size may have contributed to the statistically significant associations found in the study by Davis and Cooper. 14Ȯ 2peak normalized for BM, treadmill time, and estimated VȮ 2peak/Léger et al. were the only indices of CRF positively associated with gray matter volume.However, the associations of VȮ 2peak normalized for BM and treadmill time with gray matter volume were explained by BF%.As we also observed that VȮ 2peak normalized for LBM was not associated with gray matter volume, our results suggest that body adiposity is an important confounder for the associations between CRF and brain health outcomes in children with overweight/obesity.Previous studies in children and adults reporting positive associations between CRF and gray matter volume after controlling for several confounding factors partially supported our findings.9,38,39 However, it is important to note that none of these studies considered DXA-derived BF% in their analyses.Nevertheless, peak performance in the 20mSRT quantified by laps or running speed was not associated with gray matter volume.Gray matter plays an important role in the controlling functions related to memory, emotions, and movement.40,41 However, it also has complicated maturation-related development and associations with behavioral brain health outcomes, such as executive functions.40,41 Although our results suggest that a phenotype with high peak performance and favorable body composition may benefit brain health, further longitudinal studies with larger sample sizes clarifying the role of different indices of CRF in the gray matter volume are warranted.
In contrast to some previous findings, 42,43 we found no associations between the different indices of CRF and T A B L E 2 Associations of indices of cardiorespiratory fitness with brain health outcomes., peak oxygen uptake estimated from the 20-metre shuttle run test using the equations by Léger et al., 24 Mahar et al., 25 and Matsuzaka et al., 26 respectively.Statistically significant associations are bolded.

Hippocampal volume
hippocampal volume.Although the hippocampus is important for learning and memory functions and maybe sensitive to changes in CRF, 42,44 also some previous ies have found weak associations between CRF and hippocampal volume. 42However, the reason for these weak and nonsignificant associations is unknown.While Aghjayan et al. 42 speculated that these mixed findings could be related to the assessment of CRF, we showed nonsignificant associations between indices of CRF and hippocampal volume using a variety of CRF measures.Therefore, the association between CRF and hippocampal volume requires further clarification.
Our findings indicate that using different equations to estimate VȮ 2peak from 20mSRT performance influences the associations between CRF and brain health outcomes.Therefore, our results suggest that the direct measures from the 20mSRT, such as laps or maximal running speed, are more appropriate for examining associations between CRF and brain health outcomes.Nevertheless, if any equation is to be used in relation to brain health outcomes, our findings showed that the Leger equation provides the most consistent associations with different measures of brain health.The reason for different findings in our population may be due to the different variables in each equation.Whereas the equation by Léger et al. 24 uses only age and maximal running speed in the equation, the equations by Mahar et al. 25 and Matzuzaka et al., 26 developed in populations including mainly children with normal weight, also include body mass index.Including a measure of adiposity in the equation may introduce a larger error in the estimation of VȮ 2peak among children with overweight/obesity.
Contrary to our previous study among adolescents and young adults, 18 we found no associations between VȮ 2 at VT and brain health outcomes.Submaximal indices of CRF may be differently related to brain health outcomes in adolescents and young adults than in children.Moreover, it is also possible that different tasks used to assess brain health may influence findings.Nevertheless, absolute VȮ 2 at VT, but not relative, has been positively associated with gray matter volume in adults. 39However, the association between absolute VȮ 2 at VT and gray matter volume was weaker than the associations with the maximal indices of CRF. 39Our findings suggest that VȮ 2 at VT is less important for brain health compared to peak performance in children.
Performance in the 20mSRT and estimated VȮ 2peak/Léger et al. was positively associated with academic performance in girls but not boys.While the reason for these sex differences is unclear, girls may have better motivation towards achieving high peak performance and academic performance than boys.Moreover, we also found that better performance in the 20mSRT was associated with better academic performance and a higher estimated VȮ 2peak/Léger et al. with larger gray matter volume in children with higher BF%.These findings suggest that peak performance is more important for brain health than directly measured VȮ 2peak and that the importance of peak performance is accentuated in children with high BF%.It can be speculated that peak performance may protect against obesity-induced deterioration of brain health in children by decreasing, for example, insulin resistance and systemic low-grade inflammation and improving cerebral blood flow. 45,46However, the 20mSRT performance was more strongly associated with executive functions in children with lower BF% than their peers with higher BF%.Executive functions refer to higher-order cognitive processes essential for goal-directed behavior. 36As such, while this association is counterintuitive, the association may be related to reverse causality, as children with lower BF% and better executive functions may be more motivated and willing to run longer in the 20mSRT.However, this finding should be interpreted with caution because the association was not uniform across all measures derived from the 20mSRT.
The strengths of our study include the comprehensive assessment of CRF using laboratory-and field-based measures.Moreover, we also utilized a comprehensive assessment of brain health outcomes, including behavioral measures and structural brain imaging.However, some potential limitations should be acknowledged.The limitations of our study include the possibility that indices of CRF would be differently associated with other brain structures than total gray matter and hippocampal volume used in this study.However, a detailed analysis of brain structures was beyond the scope of this study.Gray matter and hippocampal volumes have been positively associated with academic performance. 9,47We also investigated the associations of indices of CRF with brain health outcomes in a sample of children with overweight/obesity.Whether the associations would be similar among children with lower levels of adiposity or clinical conditions is unknown.Only 66% of the children met the predetermined criteria for the maximal effort in the treadmill exercise test used to assess VȮ 2peak .Therefore, it is possible that all children did not achieve their true maximal cardiorespiratory capacity on this test.However, the results remained materially unchanged in the sensitivity analyses, which included only those children who met the predetermined criteria for maximal effort, suggesting that the primary analyses provided robust results.Furthermore, the sample size was relatively small particularly in the sex-stratified analyses.Finally, our study was cross-sectional, precluding any causal interpretations.
In conclusion, we found that better peak performance in the 20mSRT, expressed as laps and final speed achieved, positively associated with executive functions and academic performance in children with overweight/ obesity.Alternatively, directly measured VȮ 2peak normalized for either BM or LBM had weak, if any, associations brain health outcomes.Compared with the other equations, VȮ 2peak estimated using Leger and colleagues' equation 24 provided the strongest associations with brain health outcomes.We did not find any associations between brain health outcomes and submaximal indicators of CRF (e.g., VȮ 2 at VT).Finally, our results suggest that different measures of CRF cannot be used interchangeably in studies investigating associations between CRF and brain health in children.Future longitudinal studies are warranted to investigate whether our results can be replicated.Moreover, studies investigating the mechanisms how different indices of CRF influence brain health are needed.

| PERSPECTIVE
Even though previous studies have shown that higher CRF is associated with improved brain health in children, various methodologies used to assess CRF may have clouded our understanding of the importance of CRF for young people's brain health outcomes.Thus, studies using different field tests and equations to estimate peak oxygen uptake from these tests and inappropriate scaling procedures to normalize measured peak oxygen uptake have provided evidence that children with better ability to run prolonged periods may have better brain health than other children.Our findings that peak performance measured using the 20-metre endurance shuttle run test had consistent positive associations with brain health outcomes.Moreover, these results also indicate that true CRF as a measure of cardiovascular capacity to deliver oxygen and skeletal muscle aerobic capacity may not be relevant for brain health in children and that a physical ability combining endurance, motor skills, and body composition may be beneficial for brain health.These findings suggest that while CRF is positively associated with brain health in children with overweight/obesity, not all measures of CRF are associated with brain health.

AUTHOR CONTRIBUTIONS
EAH conceptualized the work, performed statistical analyses, and drafted the first version of the work; AP-F, LG-M, PS-U, CC-S, and IE-C collected the data and contributed and revised the intellectual content of the work; DRL, TJ, and ARB conceptualized the work and contributed and revised the intellectual content of the work; FBO conceptualized the work and contributed and revised the intellectual content of the work and will act as guarantor for the paper; All authors have read and approved the final version of the manuscript, and agree with the order of presentation of the authors.

Note: 1
The data are means and standard deviations.p Value for the differences between girls and boys from Student's t-test, Mann-Whitney U-test, or chi-square test.Maximal effort was defined as meeting three out of four following criteria: achieving >85% of aged-predicted maximal heart rate, a respiratory exchange ratio of ≥1.0, volitional fatigue (i.e., >8 points in the OMNI scale) and a plateau in the oxygen uptake during the last two exercise work rates (<2.0 mL/kg/min).VȮ 2peak/Léger et al. , VȮ 2peak/Mahar et al. , VȮ 2peak/Matsuzaka et al. (speed) , and VȮ 2peak/Matsuzaka et al. (laps), peak oxygen uptake estimated from the 20-metre shuttle run test using the equations by Léger et al.,24 Mahar et al.,25 and Matsuzaka et al.,26 respectively.a Medians and interquartile ranges.b Median and minimum and maximum values.Associations of indices of cardiorespiratory fitness with brain health outcomes.Data are standardized regression coefficients with their 95% confidence intervals adjusted for sex, time to peak height velocity, and parental education.VȮ 2peak , peak oxygen uptake; mL, milliliter; BM, body mass; LBM, lean body mass; VȮ 2 , oxygen uptake; VT, ventilatory threshold; 20-m SRT, 20-metre shuttle run test;VȮ 2peak/Léger et al. , VȮ 2peak/Mahar et al. , VȮ 2peak/Matsuzaka et al. (speed) , and VȮ 2peak/Matsuzaka et al. (laps) , peak oxygen uptake estimated from the 20-metre shuttle run test using the equations by Léger et al.
et al.