The mediating role of physical fitness in the relationship between 24-h movement guidelines and arithmetic fluency in adolescents: the MOVESCHOOL study

Abstract

The clustering of unhealthy movement behaviours is highly prevalent among adolescents and significantly impacts both health and academic outcomes. While evidence linking 24-hour Movement Guidelines to general academic outcomes is substantial, research specifically on arithmetic fluency remains limited. This cross-sectional study analysed 578 Spanish adolescents (aged 11–15 years) to quantify compliance with physical activity, screen time, and sleep recommendations and to explore the role of physical fitness in this relationship. Results revealed alarmingly low adherence: only 1% of adolescents meet all three health recommendations, while nearly half meet none. Results demonstrate a threshold effect, showing that adolescents who follow at least two guidelines achieve modest but significant higher scores in arithmetic fluency (B = 7.81, η²_p = .015, p = .005). Notably, this academic advantage is statistically accounted for by higher levels of cardiorespiratory fitness, which is consistent with a mediating role in all comparisons (B from 0.88 to 2.43; all 95% CI exclude zero), whereas muscular fitness does not. Findings suggest promoting combined healthy habits may be more effective for cognitive development than targeting single behaviours. Integrating movement-based strategies into school environments could enhance potentially arithmetic fluency success by fostering the cardiorespiratory health necessary for optimal arithmetic learning.

Introduction

Large-scale international studies indicate that most adolescents engage in multiple unhealthy behaviours, such as insufficient physical activity (PA), high sedentary behaviour, inadequate sleep, poor dietary patterns, or substance use, with more than half reporting a clustering of three or more unhealthy habits, which tends to increase with age and is shaped by both individual and socio-environmental factors^1,2. These lifestyle patterns have been linked not only to adverse health outcomes³ but also to detrimental effects on academic performance⁴.

Considering the combined impact of PA, sedentary behaviour, and sleep on adolescents’ health and academic outcomes, evidence-based guidelines integrating all daily movement behaviours have been developed, such as the Canadian 24-Hour Movement Guidelines (24-h MG) for Children and Youth⁵. Despite the potential benefits of adhering to the 24-h MG, global compliance remains low. Across studies, adherence rates to all three recommendations ranged from 0% to 53.6%, with 87% of articles reporting rates below 10%⁶. This trend is further reflected in Spain, where adherence among adolescents declined dramatically during the COVID-19 pandemic, dropping from 3.3% before the pandemic to 0.2% one year later, while the proportion meeting none of the recommendations increased from 28.9% to 66.3%⁷. Although more recent evidence from 2025 suggests a slight recovery, with approximately 11% of Spanish youth meeting all three guidelines, compliance remains far from optimal^8,9.

Compliance also varied by sex, with girls significantly less likely than boys to meet al.l guidelines (3.75% vs. 6.89%), although no sex differences were observed in adherence to the individual recommendations¹⁰. Age-related patterns were also evident, with adherence to all three guidelines of 11.26%, 10.31%, and 2.68% among pre-schoolers, children, and adolescents, respectively¹⁰, identifying adolescence as the most critical period for intervention. However, a significant limitation of the current evidence base is that these prevalence rates have been largely derived from self-reported questionnaires, which are subject to recall and social desirability biases⁶. Therefore, there is a need for updated research that utilises device-based measurements to accurately assess PA and sleep current adherence levels in Spanish adolescents and to determine whether the post-pandemic decline has truly been reversed.

According to a recent systematic review by Hossian et al.⁶, compliance with these guidelines is associated with a lower risk of obesity, mental health issues, and cardiometabolic problems, as well as higher physical fitness (PF). However, research correlating these guidelines with academic variables remains less abundant. Previous studies have linked compliance with the 24-h MG to better academic performance in adolescents. However, the strength of these associations varies depending on the method used to assess guideline adherence, and confounding variables such as sex. For example, in a study of Spanish adolescents, meeting all three guidelines was associated with significantly higher academic achievement compared to those meeting none¹⁰. Similarly, US adolescents in the YRBS 2019 Cycle Study who met all three guidelines were more likely to report first-class academic achievement, with higher odds for girls than boys¹¹. Additionally, Swiss adolescents who adhered to all three guidelines had higher academic achievement scores compared to those who met none¹². However, the study by Chen et al.¹³ found a positive association between adherence to the 24-h MG and academic performance in Chinese children, but not in adolescents. These relationships are theoretically grounded in the influence of movement behaviours on neurobiological and cognitive precursors. Specifically, PA, particularly when enhancing CRF, triggers growth factor cascades that strengthen synaptic plasticity in brain regions like the hippocampus¹⁴, while adequate sleep and limited screen time optimise memory consolidation and reduce the cognitive load on the prefrontal cortex¹⁵. These adaptations collectively enhance executive functions, such as working memory and inhibitory control, which are essential for processing complex information.

Given the connection between PA, sedentary behaviours and sleep, one critical academic outcome that may be affected is arithmetic fluency, which depends heavily on underlying cognitive processes. Arithmetic fluency refers to the rapid, accurate, and effortless mental calculation of sums and products for numbers from 0 to 10, including inverses, and has been widely used as a measure of mathematical ability¹⁶. Its strong association with broader cognitive performance and its predictive value for academic achievement make it an especially suitable indicator for investigating links between daily movement behaviours and adolescents’ learning outcomes^17,18.

Compliance has also been linked to higher levels of PF, including cardiorespiratory fitness (CRF) and overall muscular fitness (MF)^6,19. These fitness components, particularly CRF, have been associated with cognitive and academic performance²⁰. In contrast, evidence for MF is more mixed, with studies reporting less conclusive associations with learning outcomes²¹. Given their influence on cognition, these PF components have been hypothesised as key mediators in the relationship of movement behaviours with academic outcomes and may therefore help to explain links between guideline adherence and academic performance. Evidence from a systematic review by Visier-Alfonso et al.²² supports this mediating role as the authors reported that in most cases (six out of nine), PF (primarily CRF), exerted a significant indirect effect on the association between various forms of PA and academic performance. However, these findings were often nuanced, with some effects being subject-specific (e.g. mediation for mathematics but not for reading) or sex-dependent. Crucially, while the mediating role of PF is gaining recognition for isolated behaviours such as PA, it remains largely unexplored within the holistic framework of the 24-h MG. This represents a significant gap in the literature, as it is still unclear how the integration of PA, sedentary behaviour, and sleep collectively translates into academic benefits through PF.

Building on the above, the present study aims to: (i) quantify adherence to the international 24-h MG among Spanish adolescents; (ii) examine the associations between compliance with these guidelines and arithmetic fluency; and (iii) explore the potential mediating role of CRF and MF in the relationship between 24-h MG adherence and arithmetic fluency.

Methods

Design and participants

As part of the MOVESCHOOL study (protocol number NCT06254638)²³, a sample of 578 Spanish adolescents (aged 11–15 years; 307 boys) was selected for the present analysis based on the availability of complete and valid data for 24-h MG variables. This cohort was recruited through convenience sampling from 11 secondary schools across the Spanish provinces of Caceres and Cadiz. Participants were eligible if they: (i) were enrolled in 7th or 8th grade; (ii) had no physical disability or health condition that would restrict engagement in PA; and (iii) provided written informed consent from a parent or legal guardian. Participants also provided assent and were informed that they could withdraw from the study at any time without consequence. An a priori power analysis indicated that 432 participants were required to detect a small-to-medium effect size (f = 0.15), with an alpha level of 0.05 and a power of 0.80 for Analysis of Covariance (ANCOVA) (GPower 3.1.9.4, Düsseldorf, Germany). Regarding the mediation models, a sample size > 500 is necessary to ensure stable indirect effect estimates and a power of 0.80^24,25. Our analytical samples exceeded these thresholds, ensuring robust statistical power across all tested models.

The study was approved by the Bioethics Committee of the University of Cadiz (Cadiz, Spain), as well as the Bioethics and Biosafety Committee of the University of Extremadura (Caceres, Spain). All methods were performed in accordance with the relevant guidelines and regulations (including the Declaration of Helsinki). Furthermore, informed consent was obtained from all participants and their parent or legal guardians.

Measures

Physical activity levels and sleep time

Physical activity and sleep time were device-measured using Actigraph GT3X+ accelerometer (Actigraph Inc., Pensacola, FL, United States). Participants were instructed to affix the monitor to their non-dominant wrist and wear it for seven consecutive full days. The device was configured to capture accelerations at a sampling frequency of 100 Hz. Following the completion of the data collection period, the monitors were retrieved and returned to the research facilities for data extraction and subsequent processing. Raw sensor data were processed with the open-source R package GGIR (version 3.2.6)²⁶. The GGIR processing pipeline systematically included the following stages: (i) calibration: acceleration data were calibrated using the automatic algorithm developed by van Hees et al.²⁷; (ii) non-wear detection: non-wear time was identified based on the standard deviation and range of the raw accelerations, applying the default methodology in GGIR, which is the updated algorithm by van Hees et al.²⁸ as described here; (iii) data quality assurance: consistent with GGIR standard, potential time gaps in the accelerometer raw data were imputed, and clipping time, characterised by unusual sustained high accelerations, was detected and classified as invalid; (iv) acceleration features: key metrics derived from the raw accelerations included the Euclidean Norm Minus One G (ENMO) with negative values rounded to zero²⁹, the angle of the z-axis of the device relative to the horizontal plane, and step counts; (v) Raw acceleration data were processed and aggregated into 5-second epochs. Subsequently, PA intensity classification: PA intensity was categorised using cut-points of 35 mg for light intensity, 200 mg for moderate intensity, and 700 mg for vigorous intensity^30,31. A valid day required the device wear time at least 10 h during the awake period. For a participant’s data to be included, records had to contain at least three valid weekdays and one valid weekend day.

For this specific study, moderate-to-vigorous PA (MVPA) was utilised, which was calculated as the sum of time spent in moderate and vigorous PA categories. Consistent with the 24-h MG, participants were deemed to meet the PA guidelines if their daily MVPA accumulated at least 60 min. The sleep period time (SPT) was detected as the main nocturnal sleep episode for each recording night. The detection methodology depended on data availability: if the wear-time fraction between noon and noon surrounding each night was less than 25%, the SPT was detected using the “NotWorn” algorithm implemented in GGIR, which classifies the longest consecutive non-wear window as the sleep period time. Conversely, if the available time was 25% or greater, the SPT was determined using the HDCZA algorithm, which analyses variability in the z-axis angle throughout the day^32,33. To ensure the reliability of the 24-h movement estimates, only participants with at least three valid weekdays and at least one valid weekend day were included in the final analysis. According to the 24-h MG for children and adolescents, students were considered compliant with the sleep duration guidelines if they accumulated 9 to 11 h per day in children aged 5 to 13 years or 8 to 10 h per day adolescents aged 14 to 17 years.

Screen time

Screen time was assessed using specific items derived from the Youth Leisure-time Sedentary Behaviour Questionnaire (YLSBQ)³⁴. This instrument measured the daily duration participants dedicated to various recreational screen-based activities, encompassing television viewing, video game playing, computer use, web browsing, and mobile device or tablet utilisation (e.g., social media, WhatsApp). Participants reported their screen time separately for weekdays and weekends, selecting from the following response categories for daily time expenditure: ‘0 minutes’, ‘30 minutes’, ‘1 hour’, ‘2 hours’, ‘3 hours’, ‘4 hours’, or ‘5 h or more’. The total average daily screen time was subsequently computed as a weighted mean [(weekday screen time × 5) + (weekend screen time × 2)] / 7. Based on the 24-h MG for children and adolescents, participants were considered compliant with the recreational screen time guidelines if their time was ≤ 2 h/day.

Physical fitness

Physical fitness was evaluated across two key components CRF and MF, adhering to the standardised protocols of the ALPHA Health-Related Fitness Test Battery³⁵. All evaluations were conducted by trained researchers to ensure protocol consistency and data quality. Cardiorespiratory fitness was quantified using the 20-meter shuttle run test, which commenced at an initial speed of 8.5 km/h, with the speed incrementally increasing by 0.5 km/h per stage. Participants shuttled back and forth between two lines set 20 m apart, guided by an audio signal. The test ended if the participant stopped due to fatigue or when failed to reach the designated line before the corresponding signal sounded. Muscular fitness was assessed through two separate measures, handgrip strength (upper body) and the standing long jump (lower-body). Upper body strength was measured using a digital hand dynamometer (TKK 5101 Grip D; Takey, Tokyo, Japan). The device was adjusted according to established protocols specific to sex and hand size, as described by Ruiz et al.³⁶. Each hand was tested twice, and the mean of the best measured value for each hand, recorded in kilograms (kg), was retained for analysis. Lower-body strength was evaluated via the standing long jump. Participants performed two attempts, and the longest distance achieved, measured in centimetres (cm), was recorded. Age- and sex-specific z-scores were calculated for CRF and MF, with MF represented as a composite score averaging the standardised standing long jump and body-weight–adjusted handgrip strength scores³⁷.

Arithmetic fluency

To measure arithmetic fluency, Test 6 of the Woodcock Protocol was administered^38,39. Testing was conducted individually in quiet, controlled conditions. Participants were instructed to solve as many simple mathematical calculations as they could, out of a total of 160, correctly within a three-minute limit, emphasizing the need for both speed and accuracy. This test has demonstrated adequate validity and reliability for assessing arithmetic fluency in school-aged populations.

Covariates

Sociodemographic data included sex, age, and socioeconomic status (SES), which was measured with the Spanish-adapted version III of The Family Affluence Scale (FAS III)⁴⁰. The FAS III comprises six items assessing family purchasing power, each rated on a categorical scale. A total score ranging from 0 to 13 was calculated by summing responses across all items.

Data analysis

Descriptive statistics (mean and standard deviation) were used to characterise the overall sample, and sex differences were tested using an independent-samples t-test (boys vs. girls). Bivariate correlational analyses (Pearson’s correlation coefficient) were also conducted to examine relationships among the main study variables. In addition, descriptive statistics were used to calculate the proportion of participants meeting each of the 24-h MG recommendations (PA, recreational screen time, and sleep duration). A Venn diagram was used to visualise the overlap and the proportion of participants meeting different combinations of these guidelines. Given the small number of participants who adhered to all three recommendations, a single categorical variable was created, grouping participants into three categories for subsequent analyses: (1) meeting no recommendations, (2) meeting one recommendation, and (3) meeting two or more recommendations. To examine the association between meeting the 24-h MG (as a categorical variable) and arithmetic fluency, an ANCOVA was performed. This analysis tested mean differences in arithmetic fluency across the three guideline groups mentioned above while controlling for potential confounders (sex, age and SES). Prior to conducting the ANCOVA, all parametric assumptions were verified. Homogeneity of variances was confirmed by Levene’s test. Additionally, normality of residuals was confirmed through visual inspection of Q-Q plots, and the linearity of the covariate-outcome relationship was verified via scatterplots. All assumptions were successfully met.

Finally, to examine whether the association between adherence to the 24-h MG (X) and arithmetic fluency (Y) was mediated, separate mediation analyses were conducted for CRF (M) and MF (M), adjusting for sex, age, and SES. The categorical adherence variable (meeting no guidelines, meeting one guideline, and meeting two or more guidelines) was treated as a multicategorical predictor using indicator (dummy) coding. The “no guidelines” group was initially set as the reference category to evaluate the effects of meeting one, two, or more guidelines. To facilitate pairwise comparisons between all levels, specifically between the one and two or more guidelines’ groups, the models were re-estimated by rotating the reference category accordingly. Separate mediation analyses were conducted for CRF and MF using simple mediation (Model 4) of the PROCESS macro (version 4.2 for SPSS), testing each variable independently as a mediator. The analysis investigated the total effect of guidelines compliance on arithmetic fluency (c), the direct effect of guidelines compliance on the mediators (a path), the effect of the mediators on arithmetic fluency (b path), and the direct effect of guideline compliance on arithmetic fluency after controlling for the mediators (c’). The unstandardised indirect effect (IE), calculated as the product of the a and b coefficients (a x b), represents the portion of the total effect of guidelines compliance on arithmetic fluency that operates through the proposed mediators. To test the statistical significance of the indirect effects, bias-corrected non-parametric bootstrapping with 5,000 resamples was used to compute 95% confidence intervals (CI). This method is inherently robust as it does not assume normality of the indirect effect²⁴. Significance was established if the 95% CI did not include zero²⁴.

Sensitivity analyses were conducted using alternative operationalizations of SES, including SES as a categorical variable (low and high). The pattern of results remained consistent, with guideline adherence significantly associate arithmetic fluency and the indirect effect remaining statistically significant.

Statistical significance for all tests was set at p < .05. All statistical analyses were conducted using IBM SPSS version 25 (IBM Corp., Armonk, NY, USA).

Results

Missing data across variables ranged from 0% to 5.19%, specifically: arithmetic fluency (n = 2, 0.35%), SES (n = 1, 0.17%), CRF (n = 25, 4.33%), and MF (n = 30, 5.19%), while no missing values were present for sex or age. Consequently, a complete case analysis approach was adopted. Of the initial 578 participants full 24-h MG data, 575 were included in the primary ANCOVA models. For the mediation analyses, missing data in the mediators resulted in final samples of n = 547 for the CRF models and n = 545 for the MF models.

Sociodemographic and guideline adherence characteristics

Table 1 summarises the sample characteristics and sex differences. Boys exhibited significantly higher mean values than girls in CRF and both MF components (all p < .001), as well as for all behavioural variables except for sleep time (p = .780). Furthermore, Pearson’s bivariate correlation coefficients are displayed in the matrix, where the values below the main diagonal represent correlations for boys, and those above the diagonal indicate correlations for girls. The correlation analysis established significant relationships between the 24-h MG, fitness components, and arithmetic fluency. Moderate-to-vigorous physical activity showed a positive correlation with arithmetic fluency, while screen time was negatively correlated significant only in boys (r = .148, p < .01; r = − .147, p < .01, respectively). Cardiorespiratory fitness showed significant correlations with movement behaviours across the sample, specifically associating negatively with recreational screen time in both boys and girls, and positively with MVPA (all p < .05). Regarding academic outcomes, CRF was consistently and positively associated with arithmetic fluency in both boys and girls (all p < .05). In contrast, relative upper-body muscular strength (handgrip/weight) and lower-body muscular strength correlated significantly with arithmetic fluency only in girls (all p < .001).

Table 1.

Descriptive and correlational analysis of physical activity, sleep time, screen time, physical fitness and arithmetic fluency by sex.

	Total Sample	Boys	Girls	p		Correlations
	(M ± SD)	(M ± SD)	(M ± SD)		1	2	3	4	5	6	7
1. MVPA (min/day)	42.26 ± 22.82	51.12 ± 24.09	32.24 ± 14.85	< 0.001	—	0.045	− 0.109	0.324***	0.081	0.327***	0.113
2. Sleep time (min/day)	509.90 ± 65.48	510.62 ± 68.67	509.09 ± 61.78	0.780	0.079	—	− 0.012	0.038	0.018	0.096	0.103
3. Screen time (min/day)	238.79 ± 150.08	256.25 ± 148.47	219.00 ± 149.72	0.003	− 0.081	0.004	—	− 0.212**	− 0.082	− 0.070	− 0.099
4. CRF (stages)	3.86 ± 2.17	4.78 ± 2.26	2.80 ± 1.48	< 0.001	0.207***	0.115*	− 0.244***	—	0.271***	0.469***	0.163**
5. Upper-body muscular strength (kg/kg)	0.89 ± 0.19	0.93 ± 0.20	0.85 ± 0.16	< 0.001	− 0.072	0.115*	0.004	0.564***	—	0.421***	0.158**
6. Lower-body muscular strength (cm)	144.03 ± 28.07	155.49 ± 27.60	131.12 ± 22.45	< 0.001	− 0.007	0.062	− 0.045	0.545***	0.601***	—	0.219***
7. Arithmetic fluency	83.52 ± 21.14	87.15 ± 20.42	79.41 ± 21.21	< 0.001	0.148**	0.048	− 0.147**	0.260***	0.009	0.098	—

Notes: M: Mean; SD: Standard deviation; MVPA: moderate-to-vigorous physical activity; CRF: cardiorespiratory fitness. Correlation coefficients below the diagonal refer to boys, and those above the diagonal refer to girls. (*): p < .05, (**): p < .01 and (***): p < .001.

Figure 1a illustrates the proportion of participants meeting the individual and combined 24-h MG recommendations for the total sample, with Fig. 1b and c presenting the corresponding results for boys and girls, respectively. Overall, 45.7% of the sample met none of the 24-h MG recommendations. Among those meeting at least one recommendation, most adhered to only one (40.9%), whereas only 1.0% achieved full compliance with all three.

Fig. 1.

Proportion of adolescent adhering to the 24-h Movement Guidelines. The Venn diagram (a) refers to the total sample, (b) to boys, and (c) to girls.

Distinct sex-specific patterns emerged regarding guideline adherence. Boys were more likely to meet the PA recommendation than girls, both when met exclusively (17.3% vs. 3%) and when combined with other guidelines (13% vs. 2.9%). In contrast, girls showed higher adherence to the screen time recommendation than boys (24.4% vs. 14% considering both exclusive and combined compliance).

Association between adherence to the 24 h movements guidelines and arithmetic fluency

Overall, greater compliance with the movement guidelines was significantly associated with higher arithmetic fluency performance (p = .014, η²_p = .015; Fig. 2). The model explained 5.2% of the variance in arithmetic fluency (R² = 0.052). Post-hoc pairwise comparisons showed that students with higher adherence to the 24-h MG (meet two or more) was positively associated with higher arithmetic fluency score, with significant differences found when compared to meeting one guideline (89.82 ± 2.38 vs. 83.22 ± 1.35, p = .049) and no guidelines (89.82 ± 2.38 vs. 81.89 ± 1.28; p = .011). These differences represented small-to-moderate (d = 0.38) and moderate (d = 0.44) effect sizes, respectively. Detailed raw scores and effect sizes for all comparisons are presented in Table 2.

Fig. 2.

Differences between adolescents that met none (n = 262), met one (n = 235), met two or more (n = 78) 24-h Movement Guidelines and arithmetic fluency. Error bars represent 95% Confidence Intervals. Cohen’s d values are provided for significant pairwise comparisons as effect size estimates (* = p < .05). The analysis was adjusted for sex, age and socioeconomic status (R²= 0.052).

Table 2.

Pairwise comparisons, raw differences, and effect sizes for arithmetic fluency.

Comparison	Mean diff. (raw)	Pooled SD	Cohen´d
No meeting vs. Meeting one guideline	1.25	21.12	0.06
No meeting vs. Meeting two or more guidelines	8.99*	20.89	0.44
Meeting one vs. Meeting two or more guidelines	7.74*	20.81	0.38

Mean diff. (raw) represents the absolute point difference in arithmetic fluency; SD: Standard deviation; (*): p < .05.

No significant difference was observed between those who met one and meet none (p > .05).

Mediating effect of cardiorespiratory fitness and muscular fitness

Figure 3 presents the mediation analyses examining whether adherence to the 24-h MG predicted arithmetic fluency through CRF, while Fig. 4 shows the corresponding analyses for MF. Mediation analyses indicated that CRF significantly mediated the association between adherence to the 24-h MG and arithmetic fluency. Adherence to the guidelines was modelled as a categorical predictor with three levels (meeting no recommendations, meeting one recommendation, and meeting two or more recommendations), using meeting no recommendations as the reference category. Compared with participants meeting no recommendations, those meeting one recommendation showed a significant indirect effect via CRF (B = 0.88, BootSE = 0.41, 95% CI [0.20, 1.79]), whereas both the total effect (B = 1.27, SE = 1.89, p > .05) and the direct effect (B = 0.39, SE = 1.88, p > .05) were non-significant. A stronger indirect effect was observed for participants meeting two or more recommendations compared with those meeting none (B = 2.43, BootSE = 0.78, 95% CI [1.08, 4.12]). Although the total effect was significant in this comparison (B = 7.81, SE = 2.74, p = .005), the direct effect was no longer statistically significant after inclusion of CRF, although it approached the threshold of significance (B = 5.38, SE = 2.76, p = .052). Finally, when comparing participants meeting two or more recommendations with those meeting one recommendation, the indirect effect via CRF remained significant (B = 1.54, BootSE = 0.64, 95% CI [0.46, 2.95]). In this contrast, the total effect was significant (B = 6.54, SE = 2.76, p = .018), whereas the direct effect again approached, but did not reach, statistical significance (B = 4.99, SE = 2.74, p = .069).

Fig. 3.

Simple mediation model (n = 547) with CRF as a mediator between (a) no meeting vs. meeting one guideline, (b) no meeting vs. meeting two or more guidelines, (c) meeting one vs. meeting two or more guidelines, and arithmetic fluency. B = non standardised effect; β = standardised effect; IE = indirect effect; %Med = proportion mediated; CRF, cardiorespiratory fitness. * = p < .05; *** = p < .001. All models were adjusted for sex, age and socioeconomic status.

Fig. 4.

Simple mediation model (n = 545) with MF as a mediator between (a) no meeting vs. meeting one guideline, (b) no meeting vs. meeting two or more guidelines, (c) meeting one vs. meeting two or more guidelines, and arithmetic fluency. B = non standardised effect; β = standardised effect; MF, muscular fitness. * = p < .05; ** = p < .01; *** = p < .001. All models were adjusted for sex, age and socioeconomic status.

Muscular fitness did not significantly mediate in the association between adherence to the 24-h MG and arithmetic fluency. Specifically, the indirect effects via MF were non-significant for the contrasts of meeting one recommendation versus none (B = 0.37, BootSE = 0.27, 95% CI [-0.07, 1.00]), meeting two or more recommendations versus none (B = 0.65, BootSE = 0.43, 95% CI [-0.02, 1.63]), and meeting two or more recommendations versus one (B = 0.28, BootSE = 0.37, 95% CI [-0.40, 1.10]). Notably, the total and direct effects were significant only for comparisons involving meeting two or more recommendations (p < .05), with SE ranging from 2.72 to 2.74. Overall, these findings suggest that the association between 24-h MG adherence and arithmetic fluency operates largely independently of MF.

Discussion

This research quantifies adherence to the 24-h MG among Spanish adolescents, examines associations between guideline compliance and arithmetic fluency, and explores the potential mediating roles of CRF and MF in these associations. Although only 1% of participants met all three recommendations and one-half of the sample met none of the 24-h MG recommendations, greater compliance was significantly associated with better arithmetic performance. This association appeared to follow a threshold pattern, in which meeting two or more recommendations was necessary to observe modest but statistically significant cognitive advantages compared with meeting one or none guidelines. Crucially, the mediation models indicated that these benefits were largely channelled through CRF, whereas the association between 24-h MG adherence and arithmetic fluency appeared to operate independently of MF. While these results highlight the importance of CRF, the small effect size (η²_p = .015) suggests that the academic advantage, though present at a population level, is modest in magnitude and likely influenced by other multifaceted factors. Collectively, these findings suggest that, although high adherence is rare, this academic advantage may be partially accounted for by higher CRF.

Our findings indicate a very low adherence to the 24-h MG among Spanish adolescents, both in girls and boys, consistent with the international evidence^6,19. Specifically, studies such as USYRBS⁴¹ and the ISCOLE⁴² have reported similarly low compliance rates for all three recommendations (7% and 4.88%, respectively). Notably, the proportion of adolescents who failed to meet any of the recommendations was higher in our study (45.7%) than the global estimate by Zhao et al.¹⁹ (28.8%) and falls at the upper end of the range reported worldwide by Hossian et al.⁶(25–66%), indicating a particularly unfavourable behavioural profile. At the national level, our results appear more closely aligned with the 1% post-pandemic estimate reported by Tapia-Serrano et al.⁷ than with more recent Spanish studies reporting higher compliance of around 10%^8,9. This discrepancy may be partly explained by contextual and sociodemographic characteristics of the sample, potential residual effects of the COVID-19 pandemic on adolescent lifestyles, as well as methodological differences between studies, such as the use of device-based assessments (e.g., accelerometry in our study), which are often associated with lower estimates of guideline adherence than self-reported measures. Overall, these findings suggest that, despite signs of recovery reported in recent literature, a substantial proportion of Spanish adolescents remain far from achieving a healthy 24-h movement behaviour profile.

In the context of widespread non-adherence, our findings indicate that adolescents who meet at least two of the 24-h MG recommendations exhibit slight, albeit significant, positive association in arithmetic fluency compared to those meeting one or none. Although no previous studies have specifically examined arithmetic fluency within the 24-h MG framework, these results align closely with existing evidence on global academic performance⁶. Specifically, despite differences in samples, educational systems, and outcome measures, studies conducted across diverse contexts converge on the importance of combined movement behaviours rather than isolated guidelines. For example, Lien et al.⁴³ observed that meeting multiple recommendations, particularly those related to sleep and screen time, was associated with better self-reported academic performance, with effects varying by educational stage. Evidence from Spanish adolescents further supports this pattern, as Tapia-Serrano et al.¹⁰ showed that meeting one or more recommendations was linked to higher academic achievement across sexes, specifically, the synergy between PA and sleep duration was the unique associated with academic achievement (average score in Spanish, English, and Mathematics) and consistent in both unadjusted and adjusted models (including age, sex, and BMI). While Marciano & Camerini¹², using grades score, found that meeting multiple recommendations, especially those involving screen time and PA, was associated with higher overall academic performance. Nevertheless, Chen et al.¹³, using subject-specific grades, reported that adherence to multiple 24-h MG was favourably associated with mathematics performance in younger students, whereas these associations were largely absent in high school students, with math-related benefits observed only for meeting the screen time guideline. Importantly, although most previous studies relied on global academic indices or school grades, our use of arithmetic fluency as a domain-specific academic indicator extends this body of evidence by suggesting that the benefits of meeting multiple 24-h MG may also translate into foundational mathematical skills. Together, these findings support the notion of a threshold effect during adolescence, whereby the accumulation of at least two healthy movement behaviours appears necessary to observe meaningful advantages in math-related academic performance, beyond what is achieved by meeting a single guideline.

According to McNeil et al.¹⁶, while generic brain training or broad enrichment programmes cannot substitute for direct instruction in the fundamentals of reading and mathematics, a foundation of general health, including regular PA and adequate sleep, may support the cognitive processes necessary for learning these skills. Our results reinforce this premise, suggesting that adherence to movement guidelines provides the physiological and cognitive scaffolding that can facilitate the acquisition and automatisation of arithmetic skills. From a neurocognitive perspective, this relationship may be explained through executive functions. As noted by Geary & Moore⁴⁴, mathematical performance relies heavily on central executive capacity. Given that adherence to the 24-h MG has previously been linked to improvements in cognitive control, flexibility, and executive efficiency^6,19, these executive processes may represent one of several plausible pathways linking healthier movement behaviours to arithmetic fluency. Additionally, a recent study confirmed this link even in clinical populations with ADHD, where 24-h MG adherence related to improved inattention symptoms and performance of school⁴⁵. Since ADHD-related deficits in executive control and working memory overlap with the processes underpinning arithmetic fluency, these findings reinforce the 24-h MG framework as a critical foundation for cognitive and academic functioning in both neurotypical and clinical youth.

Our findings suggest that CRF represents a pathway through which healthy movement behaviours may relate to arithmetic fluency. Additionally, in line with previous research^6,19, adolescents who comply with more guidelines tended to have higher CRF. Cardiorespiratory fitness was, in turn, associated with better performance in arithmetic fluency. This finding is consistent with the review by Rico-González et al.⁴⁶ which showed that higher CRF is generally associated with better core executive functioning and with academic performance indicators, including mathematics and reading. The review also highlighted neurophysiological evidence supporting this link, demonstrating associations between CRF and brain structure and function relevant to cognitive performance. Importantly, the direct association between guideline adherence and arithmetic fluency was largely non-significant once CRF was accounted for, indicating that the observed benefits of adherence are primarily transmitted through CRF rather than occurring independently. In terms of magnitude, the proportion of the effect mediated by CRF ranged from 23.6% to 31.1% in the significant adherence comparisons. These findings are largely consistent with the “School in Motion” study, which reported that CRF mediated 28% of the effect on numeracy among adolescents⁴⁷. This consistent role of CRF as a primary mediator has also been observed in similar Spanish populations⁴⁸, reinforcing the idea that PF acts as a stable physiological bridge between movement behaviours and academic outcomes, although longitudinal research is required to confirm the directionality of these physiological pathways. Therefore, our findings extend the understanding to the holistic framework of the 24-h MG, suggesting that the integrated pattern of PA, sedentary behaviour, and sleep collectively contributes to higher CRF and, consequently, it may provide a physiological pathway for enhanced math-related cognitive skills.

By contrast, MF showed a more limited and inconsistent pattern. In line with previous studies^6,19, our findings indicated that although MF was directly related to arithmetic fluency, it did not significantly mediate the relationship between 24-h MG adherence and the academic outcome. While guideline adherence was associated with MF in the comparison of meeting two or more recommendations versus none, the indirect effects across all models were non-significant. This might partly reflect measurement nuances, as the handgrip and standing long jump tests evaluate different constructs despite being a standard combined procedure. This is consistent with evidence suggesting that MF may contribute to academic outcomes only under specific conditions, such as when stronger associations are observed in girls⁴⁹ or through cognitive pathways directly relevant to mathematics, such as visuospatial working memory⁵⁰. This pattern may be partly explained by differences in neurophysiological adaptations, as strength-related changes occur mainly at the spinal level with limited cortical reorganisation⁵¹. Such mechanisms may be less likely to support higher-order cognitive processes, which could explain why MF did not emerge as a consistent mediating pathway despite observable bivariate associations. This discrepancy between CRF and MF as mediators may reflect distinct neurobiological pathways. While CRF is closely linked to systemic molecular changes, such as increased cerebral blood flow and enhanced neuroplasticity through brain-derived neurotrophic factor (BDNF), neurogenesis, and angiogenesis, that support the prefrontal cortex and executive functions required for academic performance^14,52, MF adaptations are often more localized.

From a practical perspective, interventions should focus on comprehensive strategies that optimise compliance with the 24-h MG. Given that the school environment can provide universal and equitable access to adolescents, implementing multi-component school-based programmes is considered an optimal approach⁵³. In a broader sense, the practical application of these findings through school-based programmes involves cost-effectiveness considerations. While implementing movement-based strategies requires resources, such as staff training and the allocation of time, these initial investments should be weighed against the potential benefits. Our findings suggest that improving adherence to the 24-h MG could not only enhance health and fitness but also support academic outcomes like arithmetic fluency, potentially offering a favourable long-term return on investment by reducing healthcare costs and school failure. These activities could be integrated into existing school structures, whether through physical education classes, active breaks, or active commuting to school, while simultaneously promoting healthy sedentary and sleep habits. Implementing cross-curricular strategies can help overcome the common barrier of limited time availability. Special attention should be given to adolescents meeting zero guidelines, as they represent the most vulnerable group and could benefit the most from these school-based strategies aimed at CRF-enhancing activities. Crucially, to implement this ethically, schools must offer cost-free activities to ensure equitable access across all socioeconomic backgrounds, integrate movement into the school day to prevent added academic pressure, and focusing on personal progress to avoid stigmatising students with lower fitness levels.

This study presents several strengths. Physical activity and sleep were measured using accelerometry, and PF was evaluated through widely standardised field tests in a large sample. Analyses accounted for key sociodemographic factors and explored the potential mediating role of PF, providing novel insights into the links between adherence to the 24-h MG and arithmetic fluency. Moreover, extending beyond prior work, our study adds a unique perspective not merely by demonstrating fitness mediation, but by combining the specific outcome of arithmetic fluency with the comparative testing of CRF and MF as the indirect pathways underlying its relationship with movement behaviours. Nevertheless, some limitations should be noted. Firstly, while missing data followed a non-differential pattern for all primary variables, a significant difference was observed for SES, as participants with incomplete records tended to have lower scores. This was statistically addressed by including SES as a covariate in all analyses to minimize potential selection bias. Secondly, the very low proportion of participants meeting all guidelines limited the ability to fully capture the potential benefits of complete adherence. Consequently, it was not feasible to employ the 8-category classification representing all possible combinations of recommendations, as the resulting cell sizes would have been too small to allow meaningful comparisons. Future research should investigate if certain subgroups require tailored approaches to overcome specific barriers to adherence. Thirdly, the use of self-reported screen time, unlike the device-measured for sleep and PA, is susceptible to recall and social desirability biases, which may have affected the accuracy of the 24-h MG adherence classification. Fourthly, the cross-sectional design of the study limits the ability to establish causality or temporal precedence. Crucially, reverse causation cannot be ruled out, as adolescents with higher academic achievement might be more likely to adopt structured, healthy movement routines. Consequently, the mediation analysis should be interpreted as exploratory. Moreover, the study did not account for psychological or emotional factors, such as stress or mental fatigue, nor potential residual confounders including baseline cognitive ability, pubertal status, previous academic performance, and parental involvement, which may influence both movement behaviours and academic performance, particularly in adolescents. Future serial mediation models (core executive functions and processing speed), sex-stratified analyses and longitudinal studies are needed to confirm these findings and further clarify the mechanisms linking movement behaviours, PF, and academic outcomes. In this sense, randomised controlled trials are essential to determine if interventions aimed at improving guideline compliance and fitness effectively lead to causal improvements in arithmetic fluency. Furthermore, multi-country replication studies across different educational systems and cultural contexts, potential threshold and research examining whether these findings generalise across other fitness components, socioeconomic groups, baseline achievement levels or different academic outcomes, would broaden our understanding of these relationships.

Conclusions

Our study provides new insights into the cognitive benefits of adhering to the 24-h MG in Spanish adolescents. Despite generally low compliance, adolescents meeting two or more recommendations demonstrated modest but statistically significant higher arithmetic fluency, suggesting that maintaining an integrated pattern of PA, sleep, and limited screen time supports mathematical performance. Mediation analyses indicated that CRF, but not MF, plays a key role in linking guideline adherence to arithmetic fluency, highlighting CRF as a particularly relevant pathway for academic-related cognitive skills. Collectively, these findings emphasise the importance of promoting CRF-enhancing PA to maximise the cognitive benefits of healthy movement behaviours. Moreover, future longitudinal studies are needed to confirm these relationships over time and to explore the causal mechanisms underlying the observed associations.

Author contributions

Conceptualization, D.S-O., and A.G-C.; methodology, D.S-O., D.C-M., and A.G-C.; formal analysis, F.M-A., D.S-O., and A.G-C.; investigation, F.M-A., D.S-O., M.V-S., D.C-M., J.M., and A.G-C.; data curation, F.M-A., and D.S-O.; writing—original draft preparation, F.M-A.; writing—review and editing, F.M-A., D.S-O., M.V-S., D.C-M., J.M., and A.G-C.; visualization, F.M-A., D.S-O., and A.G-C.; supervision, D.S-O., D.C-M., J.M., and A.G-C.; project administration, D.S-O., D.C-M., and A.G-C.; funding acquisition, D.S-O., D.C-M., and A.G-C. All authors have read and agreed to the published version of the manuscript.

Funding

This publication is part of the R&D&i project PID2022-137450OA-I00 funded by MCIN/AEI/10.13039/501100011033 and by ERDF/EU; and of the grant CNS2022-135604 funded by MCIN/AEI/10.13039/501100011033 and “European Union NextGenerationEU”. F.M-A receives funding from the University of Cadiz’s Own Plan [PU/EPIF-FPI-CT/CP/2023-046]. Additionally, this work was supported by the “Program to promote and boost research activity” of the University of Cadiz, which granted a mobility aid to the author F.M-A. and A.G-C.

Data availability

The datasets supporting the findings of this study are available in Open Science Framework repository at https://doi.org/10.17605/OSF.IO/8WNFM. The analysis code and detailed protocols are available from the corresponding author upon reasonable request.

Declarations

Competing interests

The authors declare no competing interests.