Active breaks as a physical education strategy to develop attention, concentration and attentional variability in higher education Authors: José Luis Ubago-Jiménez, Gabriel González-Valero, Félix Zurita-Ortega, Eduardo Melguizo-Ibáñez, José Manuel Alonso-Vargas Published: 2026 ABSTRACT AbstractBackgroundActive breaks are an effective tool for introducing physical activity into the classroom and improving students’ physical and mental health. This study aims (a) to evaluate changes in attention, concentration and attentional stability levels before and after the implementation of an intervention programme based on active breaks and (b) to analyse the associative relationships between attention, concentration and attentional stability before and after the intervention using a repeated-measures path analysis model.MethodThis is a quasi-experimental study with a single-group pre-test and post-test design without a control group. The sample consisted of 224 university students (20.57 ± 2.07). The d2 Attention Test was used to measure overall effectiveness, concentration, and attentional variability. The intervention lasted 15 weeks and included a weekly 10-minute active break during theory classes with physical-cognitive activities, choreographed boxing, and movement games.ResultsFindings showed improvements in attention, concentration and attentional stability following the intervention. The path analysis model also revealed strengthened associations between concentration, attentional stability and overall effectiveness after the intervention, alongside changes in the reciprocal relationship between concentration and attentional variability.ConclusionsActive breaks may represent a promising strategy in initial teacher training to optimize cognitive efficiency and promote more autonomous cognitive functioning.Keywords: Attentional variability, Concentration, Executive functioning, Classroom-based intervention, University students, Teacher education Introduction Due to the challenges of the 21st century, educational centers are central to the formation of citizenship, as these institutions are where young people develop their skills and abilities [1]. Recent research highlights the fundamental role of educational institutions in transforming knowledge into skills and abilities that can be applied to different situations throughout the teaching-learning process [2]. One of the most pressing issues observed in different educational institutions is the high level of sedentary behavior [3]. It is noteworthy that sedentary involvement during the school day exceeds 90% of the time [3]. The scientific community emphasizes the need to implement organizational changes in educational institutions that promote the incorporation of physically active methodologies during the teaching-learning process [4]. Physically active methodologies are innovative teaching tools that integrate physical movement into the teaching-learning process [5]. It has been shown that these types of active methodologies promote students’ physical, emotional, and cognitive development [5], demonstrating a close temporal connection between physical activity and different academic tasks [6]. This does not necessarily imply that physical activity and academic tasks are performed simultaneously, but rather that they can be performed serially or sequentially [6]. This type of activity is based on the theoretical framework of embodied cognition [7], which establishes that the environment and the human body are related in cognitive processes, with the mind integrated into the body’s own sensorimotor systems [7]. Based on this, training future teachers in physically active methodologies not only helps to enrich their teaching repertoire, but also prepares them to create effective, dynamic, and inclusive learning environments [8]. The application of these types of methodologies has been defined by university students themselves as useful and interesting [9] however, a low level of applicability of these types of methodologies to the university classroom has been noted [10]. The relationship between physically active methodologies and active breaks is close and complementary, as both integrate body movement into the educational context to optimize the teaching-learning process [11]. Active breaks are defined as short breaks from physical activity during the school day [11]. The relationship between active breaks and the teaching-learning process is supported by various studies [12, 13]. The application of this methodology promotes a more positive and receptive environment for content assimilation [12, 13]. Intercalating moments of physical activity with training processes promotes greater willingness by reducing mental fatigue [11]. It has also been claimed that active breaks act as a neuro-educational tool that stimulates executive functions such as working memory, planning, inhibitory control, and attention [13]. Attention is defined as the cognitive system’s ability to select relevant information and focus mental resources on specific stimuli [14]. This cognitive function is influenced by concentration, which is defined as maintaining attention on a task for an extended period of time, filtering out irrelevant stimuli [15]. It is also influenced by attentional stability, which refers to the consistency with which attentional performance can be maintained without major fluctuations [15]. During the teaching-learning process, attention is essential as it allows students to solve complex tasks and resist distractions [15]. Without adequate sustained attention and attentional stability, information processing is compromised [14], negatively affecting the comprehension, retention, and application of acquired knowledge [14]. Various studies have concluded that there is a relationship between active breaks and concentration and attentional performance [11, 13]. Theoretical framework This study is grounded in the theoretical perspectives of embodied cognition and cognitive load theory. From the embodied cognition perspective, cognitive processes are closely linked to bodily movement and sensorimotor interaction with the environment [7]. Learning is therefore not understood as an exclusively mental processes, but rather as an experience shaped by the interaction between body, cognition and context. Physically active methodologies may enhance cognitive engagement by integrating movement into learning processes facilitating attention activation and information processing [7, 11]. Cognitive load theory suggests that working memory has a limited processing capacity during learning task [11]. Prolonged theoretical instruction may increase cognitive fatigue and reduce attentional fatigue and reduce attention efficiency. Active breaks have been widely recognized as an effective strategy for reducing sedentary behavior and promoting health in educational settings [11]. Scientific evidence in higher education remains limited, particularly regarding their effects on factor such as concentration, attentional stability and attentional stability [9]. University students are frequently exposed to prolonged periods of sustained cognitive demand during lectures, which may contribute to cognitive fatigue and fluctuations in attentional performance [2, 9]. From a theoretical perspective, attention is multidimensional construct in which concentration reflects sustained and efficient focus [14, 15]. Likewise, attentional variability indicates fluctuations in cognitive performance, and both may influence overall attentional variability efficacy [14]. In this regard, active breaks that combine physical and cognitive demands may improve attentional functioning not only through direct improvements, but also by strengthening the relationships between attentional processes. The current research adopts a multidimensional understanding of attention. Concentration reflects sustained and selective attentional focus, whereas attentional variability represents fluctuation in attentional performances across tasks [16]. Lower attentional variability is interpreted as greater attentional stability [16, 17]. Based on this framework, active breaks may contribute not only to improvements in attentional outcomes, but also to modifications in the associative relationships among attentional processes. This study presents the following objectives: To compare levels of attention, concentration, and attentional stability before and after implementing the intervention program based on active breaks. To compare levels of attention, concentration, and attentional stability before and after implementing the intervention program based on active breaks. To design and adjust path analysis model that examines the associative relationship between concentration and attentional stability on the degree of attention. To design and adjust path analysis model that examines the associative relationship between concentration and attentional stability on the degree of attention. To analyse the variations in the associative relationships between concentration and attentional stability on the degree of attention after implementing an active break program. To analyse the variations in the associative relationships between concentration and attentional stability on the degree of attention after implementing an active break program. Method Study design and role of the researches The research presents a quasi-experimental pre-test post-test design with a single intervention group without a control group. The study also incorporates an explanatory relational modelling component aimed at examining associations among attentional indicator across time points. Given the absence of control group, this design allows examination of pre-post changes and associations between variables. This design does not permit strong causal inferences regarding the effects of the intervention. Observed changes should therefore be interpreted cautiously, as alternative explanations such as maturation, testing effects or uncontrolled contextual influences cannot be ruled out. The investigators were involved in the design, coordination and monitoring of the intervention programme. The active breaks were implemented by the course instructor, who followed previously standardized protocol designed by the research team to ensure consistency across sessions. Researchers were not involved in participants’ academic assessment and participation in the study had no influence on course grading or academic evaluation. Given the quasi-experimental nature of the study, the researchers acknowledged the potential influence of contextual and interpersonal factor during implementation and interpretation of the findings. To minimize procedural variability, the same instructor conducted all intervention sessions and standardized instruction were applied through the programme. Participants For reasons of convenience, recruitment was carried out in groups. The sampling frame consisted of students enrolled in the second year of the primary education degree program who were taking the course Teaching Physical Education in Primary Education. The inclusion criteria were: (a) being enrolled in a university degree program in primary education; (b) being enrolled in the course Teaching Physical Education in Primary Education; (c) being between 18 and 25 years of age and (d) having given informed consent to participate voluntarily in the study. The exclusion criteria for this study were (a) providing incomplete or internally inconsistent responses and (b) current use of medication related to diagnosed psychological disorders. Initially, 243 responses were collected, but 17 participants dropped out of the study and two did not complete one of the tests (pre-test or post-test). Thus, the final sample consisted of 224 students (mean age = 20.57; standard deviation = 2.065). In terms of gender distribution, 120 were male (53.6%) and 104 were female (46.4%). This number of participants represented 27.6% of the total number of students enrolled in the Primary Education degree program during the corresponding academic year (according to institutional estimates). Although the selection was carried out at a single institution, it was non-probabilistic and voluntary. Instruments A sociodemographic questionnaire was used to collect data on gender (male/female) and age. The d2 attention questionnaire was used to collect data related to attention. The original version of this instrument was developed by Brickenkamp and Zillmer [16]. Due to the characteristics of the sample, the version by Seisdedos-Cubero [17] was used. This instrument measures the ability to selectively attend to certain relevant aspects of a task while accurately ignoring irrelevant elements. The instrument consists of 14 lines with 47 characters, giving the subject 20 s to respond to each line. These stimuli include the letters “p” and “d”. These are accompanied by one or two lines at the top or bottom of the letter. The task consists of carefully reviewing each line and marking all letters “d” that have two lines (either both at the top, both at the bottom, or one at the top and one at the bottom). These elements are considered relevant. All other combinations are considered irrelevant. The questionnaire is evaluated using the following dimensions: Total responses (TR) refers to the number of items attempted in the 14 lines; Total correct answers (TA) is the number of relevant common items; Omissions (O) indicate relevant items that were attempted but not marked; Commissions (C) are irrelevant items that were marked. These values are used to calculate the three dimensions of the questionnaire: Total Effectiveness (TOT) shows attentional control, reflecting the relationship between the speed and accuracy of the subjects. The Concentration Index (CON) represents the balance between speed and accuracy. The Variation Index (VAR) reflects the difference between the maximum and minimum elements achieved in the line. In this study, overall attentional performance refers to the d2 TOT index. Concentration refers to the CON index, reflecting focused and efficient processing. Attentional variability refers to the VAR index, indicating fluctuations in performance, where lower values indicate greater attentional stability. The selection of the d2 Attention Test was based on its suitability for assessing multiple dimensions of attentional performance withing a single standardized task. The test provides indices of processing speed and accuracy, focused attention and inhibitory control and variability in attentional performance. Therefore, the d2 test was considered an appropriate instrument for capturing both general and specific attentional changes associated with the intervention. Two indices were used to assess the internal consistency of the instrument: Cronbach’s alpha (α) and McDonald’s omega (ω). The main dimensions of the questionnaire were evaluated: Total responses (TR), Total correct answers (TA), Omissions (O), and Commissions (C). The values obtained were as follows: TR: (α = 0.97; ω = 0.90); TA (α = 0.92; ω = 0.95); O (α = 0.92; ω = 0.87) and C (α = 0.77; ω = 0.72). Procedure First, an action protocol was drawn up in accordance with the regulations established by the Ethics Committee for Research Involving Human Subjects at the University of Granada (3132/CEIH/2023). The study was conducted in line with the ethical principles of the Declaration of Helsinki (2024). Ethical approval was obtained prior to data collection and all procedures complied with current data protection regulations, specifically Regulation (EU) 2016/679 (General Data Protection Regulation, GDPR). Participants were fully informed about the objective and procedures of the study and provided written informed consent prior to participation. Confidentiality and anonymity were guaranteed and data were used exclusively for scientific purposes. Regarding the intervention, the programme was implemented between September and December 2025 over 15 weeks. It consisted of one weekly active break integrated into a 2-hour theoretical session of the Physical Education Teaching course. Approximately halfway through each session, a 10-minute active breaks were introduced. After it, the theoretical instruction continued. The intervention was delivered by the same professor throughout all sessions to ensure consistency in implementation, following previously established and standardized protocol for all active break activities. Although no formal fidelity checklist or external monitoring systems was used to assess the implementation adherence, the use of a single instructor and a structured protocol was intended to maintain consistency across the 15-week program. This should be considered a methodological limitation, as implementation fidelity was not objectively verified. Six active break modalities were designed and implemented, grouped into three categories according to their primary characteristics: cognitive-motor activities, aerobic activities and playful movement activities. Across the 15-week intervention, these modalities were alternated during weekly 10-minute active breaks embedded within theory classes. Cognitive-motor activities included multiple-choice questions related to course content and multiple-choice questions related to general knowledge. In both modalities, students answered short questions and performed brief physical exercise contingent on their responses, combining cognitive stimulation with movement. Aerobic activities included choreographed boxing and swimming stimulation in the classroom. These activities involved continuous or coordinated movement designed to promote physical activation, motor coordination and moderate aerobic effort. Playful movement activities included rock-paper-scissors in motion and movement-based interactive dynamics designed to encourage spontaneous movements, social interaction and motivational engagement. Data analysis A combination of analytical techniques was used to address the different level of analysis in the study. Paired-sample t-test were conducted to examine pre-post differences in attentional variables. Effect sizes were calculated to quantify the magnitude of change. Path analysis model was employed to examine the associative structure among attentional variables and how these associations changed across measurement points. Comparability ckecks across time were applied to ensure consistent measurement of the observed variables across pre-test and post-test. Finally, common-method variance was assessed to control for potential bias associated with self- and test-based measurement within the same sample. Statistical and inferential analysis The IBM SPSS 25.0 statistical package was used to perform the statistical analysis of the results. For the study of normality, the distribution of differences between the values obtained in the post-test and pre-test for each of the variables analysed was evaluated. For a normal distribution, skewness values should be between − 1.5 and 1.5 [18–21]. Kurtosis values should be between − 3 and 3 [18–21]. Because the variables showed a normal distribution, parametric tests were considered. The paired t-test was used to analyse the variations between the pre-test and post-test measurements. The significance level was set at p < 0.05. To complete the inferential analysis, the absolute difference between the post-test and pre-test (Δ) was calculated as an indicator of the change produced by the intervention. The effect size was also calculated using Cohen’s d [22]. This statistic was calculated to estimate the magnitude of the change observed between the pre-test and post-test values. Although this metric was originally developed to compare two independent groups [22], its use in intra-subject designs is widely supported by specialized scientific literature [23, 24]. Cohen’s d was retained to facilitate interpretation of the magnitude of pre-post changes and comparability with previous intervention studies in this filed. Although repeated-measures-specific alternatives exist, Cohen’s d provides a standardized and widely recognized estimate of practical significance for within-subject changes [22]. The effect size was calculated based on the standardized mean difference for repeated measurements, following recommendations for within-subject designs [23, 24]. The effect size classification levels have been as follows: null (≤ 0.20), small (0.21–0.49), moderate (0.50–0.79), and large (≥ 0.80) [22]. The observed statistical power (1 – β) was also calculated using the effect size and sample size, allowing the probability of detecting a real effect with the available sample to be estimated [25, 26]. 95% confidence intervals were reported for each Δ, providing a plausible range of observed effects and facilitating statistical interpretation of the results [25]. Structural modelling and comparisons across time points A repeated-measures path analysis model was specified and estimated in IBM AMOS 23.0 to examine the associations between the study variables and compare these associations after applying the intervention program. The observed changes in path model parameters should be interpreted as changes in estimated associations across time rather than evidence of causal restructuring. Given the quasi-experimental single-group design, these differences may reflect a combination of intervention-related effects, measurement variability, practice effects or model instability. Therefore, the path modelling results provide descriptive evidence of changes in the relational structure among attentional variables, but do not allow for causal inference regarding the impact of active breaks on this relationship. The fit indices obtained indicate that the proposed structural model showed an acceptable-to-excellent fit to the observed data. Specifically, the CFI and TLI values suggest adequate comparative fit [27, 28], while the low RMSEA and SRMR values indicate small residual discrepancies between the observed and estimated covariance matrices [27, 28]. Following established recommendations [27, 28], the comparative fit index (CFI) and Tucker Lewis index (TLI) were used as incremental measures of fit. The root mean square error of approximation (RMSEA) and the standardized root mean square residual (SRMR) provided information on absolute fit and residual-based discrepancies [29, 30]. CFI and TLI values ≥ 0.90 are interpreted as an acceptable fit [29, 30]. When the CFI or TLI value is ≥ 0.95, it is considered an excellent fit [29, 30]. RMSEA and SRMR values ≤ 0.08 denote an acceptable fit [29, 30], while a value ≤ 0.05 represents an excellent fit [29, 30]. Table 1 presents the fit values. Table 1Analysis of adjustment indicesModelχ²dfCFITLIRMSEASRMRTotal sample1.35830.9210.9140.0390.021Pre-test1.12230.9840.9810.0180.025Post-test1.02730.9900.9630.0290.019 Analysis of adjustment indices Assumptions and estimates Normality at the item level was preliminarily inspected using skewness and kurtosis, with values < 2, which is consistent with conventional thresholds [20, 21]. To address the multivariate normality assumption, Mardia coefficients (skewness and kurtosis) were calculated. Models were estimated using maximum likelihood and bootstrapped standard errors (2000 resamples) were obtained to improve robustness against deviations from normality. Reliability and internal consistency The reliability of the scale was assessed using Cronbach’s α and McDonald’s ω, with all coefficients ≥ 0.70. This indicates adequate internal consistency [31, 32]. In addition, composite reliability (CR) and average variance extracted (AVE) were calculated from standardized loadings (criteria: CR ≥ 0.70; AVE ≥ 0.50) to document convergent validity; discriminant validity was examined using Fornell-Larcker and HTMT (HTMT < 0.85). Comparability checks across time Procedures to assess comparability across time were conducted sequentially to examine whether the observed variables were measured consistently across pre-test and post-test assessment [28, 33]. Decisions were based in ΔCFI ≥ − 0.010 and ΔRMSEA ≤ 0.015 relative to the least restricted model [34]. The baseline comparability model showed acceptable fit at both time points (CFI ≥ 0.900; RMSEA ≤ 0.070), supporting consistency in the pattern of association across measurements. Subsequently equality constrained models presented only negligible changes in model fit, suggesting acceptable comparability of measurement across time. Full fit indices are presented in Table 2. Table 2Assessment of comparability across timeSequential comparability modelTime pointχ²dfCFIRMSEAΔCFIΔRMSEABaseline comparability modelPre-test12.45380.9540.042--Baseline comparability modelPost-test14.10680.9470.048--Equality-constrained model 1Pre-test16.204100.9530.043−0.0010.001Equality-constrained model 1Post-test17.051100.9510.044−0.0020.001Equality-constrained model 2Pre-test20.303120.9500.045−0.0020.002Equality-constrained model 2Post-test21.852120.9480.046−0.0030.001 Assessment of comparability across time Given the repeated-measures designs, these comparability checks were used to support the interpretation that observed changes reflected temporal variation in performance rather than shifts in measurement properties. Fit changes are interpreted descriptively. They should not be taken as evidence of strict longitudinal invariance. The equality-constrained comparability step showed negligible changes in model fit (ΔCFI ranged from – 0.001 to – 0.002; ΔRMSEA ≤ 0.001), supporting acceptable comparability of measurement across time. Similarly, the subsequent comparability step presented only mininmal changes in fit indices (ΔCFI up to – 0.003; ΔRMSEA ≤ 0.002), supporting consistency of measurement across time points. Theoretical representation of associative pathways in the structural model The theoretical model consists of two exogenous variables (Concentration Index and Variation Index) and one endogenous variable (Total Effectiveness). For the endogenous variables, an associated error term (e1) has been included to represent the unexplained variance. The theoretical model distinguishes between unidirectional and bidirectional associative relationships. Unidirectional associative relationships are observed between the concentration index and total effectiveness. There is also a unidirectional associative relationship between the variation index and total effectiveness. Reciprocal associative relationships exist between the variation index and the concentration index. All variables were treated as continuous and calculated as mean scores for their respective items. Figure 1 shows the conceptual model and the direction of the structural paths. . Results Descriptive statistics for all variables, along with the correlation matrix, are presented in Table 3. Skewness ranged from approximately − 1.5 to 1.5 and kurtosis from − 3 to 3, which is consistent with conventional criteria for approximate normality [20, 21]. Multivariate normality was also assessed using Mardia’s coefficients. Considering the distribution profile and the use of bootstrapped standard errors (2000 resamples), the maximum likelihood estimation for SEM was maintained [28, 35]. Table 3Mean values, standard deviations, skewness, and kurtosis of the variables before and after applying the intervention programMpreSDpreSkewpreKurtpreMpostSDpostSkewpostKurtpotTotal efectiveness286.5454.152− 0.108− 1.083350.3974.376− 0.031− 0.762Concentration index103.0429.2830.9250.678127.1434.2250.636− 0.193Variation index12.183.9540.588− 0.24719.216.0810.9810.311 Mean values, standard deviations, skewness, and kurtosis of the variables before and after applying the intervention program Table 4 presents the comparative analysis performed using the t-test for related sample. In terms of overall effectiveness, the intervention based on active breaks has significantly increased students’ overall performance (286.54 ± 54.152; 350.39 ± 74.376; p < 0.05). For the estimation of the observed effect, a value of d = 0.981 95% CI [0.822; 1.140] was obtained. Continuing with the quality of care, the intervention program has significantly increased students’ concentration levels (103.04 ± 29.283; 127.14 ± 34.225; p < 0.05). Regarding the estimation of the effect, a value of d = 0.757 95% CI [0.608; 0.906] was obtained. The intervention was associated with a significant increase in attentional stability (12.18 ± 3.954; 19.21 ± 6.081; p < 0.05), indication reduced performance variability across tasks. The effect size estimate was d = 0.834 95% CI [0.608; 0.986]. Table 4Comparative analysis of the intervention and sample distributionM ± DTΔ [95% CI]t p d [95% CI]1 − βTotal efectiveness -Pre286.54 ± 54.15263.85 [29.92; 97.78]3.85< 0.050.981 [0.822; 1.140]0.97Total efectiveness -Post350.39 ± 74.376Concentration index -Pre103.04 ± 29.28324.10 [9.69; 38.51]3.42< 0.050.757 [0.608; 0.906]0.93Concentration index -Pre127.14 ± 34.225Variation index -Pre12.18 ± 3.9547.03 [3.04; 11.03]3.60< 0.050.834 [0.608; 0.986]0.95Variation index -Post19.21 ± 6.081Mean value (M); Standard deviation (SD); Cohen’s d (d); Significance value (p); Statistical power size (1 – β) Comparative analysis of the intervention and sample distribution Mean value (M); Standard deviation (SD); Cohen’s d (d); Significance value (p); Statistical power size (1 – β) The results obtained before and after applying the intervention program are shown below (Table 5). Regarding the association between the variation index (attentional stability) and overall effectiveness, statistically significant differences were found (p < 0.05), with an increase in the standardized coefficient from βpre = 0.008 to βpost = 0.472. In relation to the effect of the concentration index on overall effectiveness, statistically significant changes were also observed (p < 0.05), with an increase int the association from βpre = 0.598 to βpost = 0.699. Finally, concerning the reciprocal association between the variation index and the concentration index, statistically significant differences were identified (p < 0.05), presenting a decrease in the standardized coefficient from βpre = 0.413 to βpost = 0.245. Table 5Analysis of associative relationships before and after implementing the intervention programPathMomentSECR p β (B)Variation index ◊ Total efectivenessPre-test0.7880.1470.0480.008 (0.116)Post-test0.44712.9230.472 (0.447)Concentration index ◊ Total efectivenessPre-test0.09713.3710.0150.598 (0.097)Post-test0.07916.3650.699 (0.079)Variation index ◊ ⇓ Concentration indexPre-test7.5925.7010.0290.413 (0.592)Post-test14.2863.5580.245 (0.986)B = unstandardised estimate; SE = standard error; CR = Analysis of associative relationships before and after implementing the intervention program B = unstandardised estimate; SE = standard error; CR = Discussion The findings of this study indicate that a four-month active break program was associated with improvements in attentional stability, concentration and attentional performance in university students. Changes were observed in the structural associations among these variables following the intervention. Given the large sample size, statistical significance should be interpreted alongside effect sizes, which provide a more meaningful estimate of the practical relevance of the observed changes. The results support previous evidence suggesting that physically active methodologies may be associated with improvements in cognitive functioning and attentional regulation in educational settings [36]. From the perspective of Cognitive Load Theory [37], attentional performance is constrained by the limited capacity of working memory. When cognitive load is high during prolonged academic instruction, fewer attentional resources remain available for sustained focus and inhibitory control [38, 39]. Prolonged exposure to theory-based classes may overload attentional resources, leading to reduced concentration and increased attentional variability. Alternating periods of academic engagement with brief movement opportunities may help optimize attentional readiness, facilitate task re-engagement and support cognitive efficiency during prolonged classroom sessions [11, 40]. The observed pre-post improvements in attention and concentration are consistent with research indicating that active breaks may positively influence executive functioning and sustained attentional performance [9, 41]. The findings also suggest that active breaks may be associated with greater attentional stability. Previous studies have shown that movement-based interruptions can contribute to reducing fluctuations in attentional performance and support a more stable attentional pattern [14, 42]. A common mechanism underlying these findings may be the reduction of mental fatigue and the stabilization of attentional control [11, 43]. It may explain the improvements in concentration, attentional consistency and overall effectiveness observed after the intervention. From an educational perspective, these processes may be relevant for tasks requiring sustained executive attention, such as problem solving, reading comprehension and information processing [11, 6]. Regarding the structural model, stronger associations were observed between the variation index and overall effectiveness, as well as between concentration and overall effectiveness. These findings may suggest a strengthening in the functional relationship among attentional components following the intervention. Prior evidence has indicated that physically active methodologies may support cognitive control, reduce attentional fatigue and promote more efficient coordination among attentional processes [38, 40, 41, 46, 47], Likewise, the reduction observed in reciprocal association between concentration and attentional variability may reflect a more autonomous and efficient attentional profile, whereby concentration may rely less on compensating for unstable attentional fluctuations. These findings suggest that active breaks may support cognitive efficiency through complementary mechanism involving physical activation, attentional stabilization and reduced mental fatigue. Although the research objectives were met, it should be noted that this study is not without limitations. The quasi-experimental design with a single group pre-test/post-test approach should be considered when interpreting the findings. Although significant improvements were observed after the intervention, the absence of a control group limits the ability to attribute these changes exclusively to the active breaks programme. Therefore, causal interpretations should be made with caution. Improvements observed from pre-test to post-test may be associated with the active break program, but they cannot be interpreted as definitive evidence that the intervention alone caused these changes, since alternative threats to internal validity may have contributed. Despite this limitation, the present findings provide preliminary evidence supporting the potential benefits of active breaks in higher education and highlight the need for future studies employing controlled or randomized designs to confirm these effects. The sample consisted of 224 students from different class groups within a single institution recruited through a convenience sampling approach. Participation was voluntary and based on informed consent and no comparison data were available for non-participants. Although attrition was minimal, potential bias related to self-selection and dropout cannot be completely ruled out. Therefore, the generalisability of the findings is limited. Results should be interpreted specifically within the context of pre-service teachers in this institution. Finally, another limitation of this study is that full longitudinal structural equation model was not applied. Therefore, it should be regarded as exploratory and interpreting with caution. Regarding future perspectives, it would be advisable to include control groups with randomized participation. This would allow for causal validation of the observed effects. Future randomized controlled studies are needed to confirm whether these observed improvements are attributable specifically to the intervention. It would also be advisable to analyse whether frequency or duration produces an incremental or different effect on cognitive variables. Another possible suggestion would be to analyse sustainability and teaching implementation. This would allow for an evaluation of the practical feasibility of integrating active breaks into the curriculum, as well as the necessary training of teachers for their effective application. Conclusions This study evaluated changes associated with a 15-week active break program in university students enrolled in Primary Education. Significant pre-post improvements were observed in attentional performance, sustained concentration and attentional stability following the intervention. In addition, changes were observed in the structural associations between the variation index, concentration index and attentional effectiveness, suggesting a potentially more robust attentional profile after the program. The reduction observed in the reciprocal association between concentration and variability may tentatively reflect a pattern compatible with greater attentional efficiency. These findings suggest that active breaks may contribute to improving attentional functioning, facilitating attention, comprehension and information processing in higher education. However, given the quasi-experimental single-group design and the absence of a control group, these findings should be interpreted with caution and no definitive causal conclusions can be drawn regarding the exclusive effects of the intervention. From an educational perspective, the results support the potential usefulness of incorporating active breaks into initial teacher training programs as a sustainable and inclusive pedagogical strategy aligned with the principles of meaningful learning and neuroeducation. Nevertheless, randomized controlled studies are needed to confirm whether the observed improvements are specifically attributable to the intervention. Acknowledgements This article is related to Project PID2024-157420NB-I00 (Aprendizaje en Movimiento y Rendimiento Académico. Nuevo Modelo Pedagógico en Educación Física basado en Descansos Activos y Resiliencia) belonging to the State Programme for Research and Experimental Development, part of the 2024–2027 State Plan for Scientific, Technical and Innovation Research of the Ministry of Science, Innovation and Universities of the Government of Spain. This study is linked to the project entitled “Descansos activos en la enseñanza superior para la mejora de la atención y rendimiento académico” which is funded under the University of Granada’s 2023 Internal Research Plan (PPJIA2023-124). Authors’ contributions Conceptualization, E.M.I. and J.L.U.J.; methodology, F.Z, O. and J.M.A.V.; software, G.G.V and E.M.I.; validation, F.Z.O. and G.G.V.; formal analysis, E.M.I. and J.L.U.J.; investigation, J.L.U.J. ; resources, F.Z.O.; data curation, G.G.V. and J.M.A.V.; writing—original draft preparation, G.G.V. and E.M.-I.; writing—review and editing, E.M.I. and J.L.U.J; visualization, J.L.U.J ; supervision, G.G.V. and J.L.U.J.; project administration, F.Z.O and J.L.U.J.; funding acquisition, J.M.A.V. All authors have read and agreed to the published version of the manuscript. Institutional review board statement The study was conducted in accordance with the Declaration of Helsinki and approved by the Research Ethics Committee (3132/CEIH/2023); the approval date was 19 January 2023. Informed consent Informed consent was obtained from all subjects involved in the study. Funding This article has been funded by Project PID2024-157420NB-I00 (Aprendizaje en Movimiento y Rendimiento Académico. Nuevo Modelo Pedagógico en Educación Física basado en Descansos Activos y Resiliencia) belonging to the State Programme for Research and Experimental Development, part of the 2024–2027 State Plan for Scientific, Technical and Innovation Research of the Ministry of Science, Innovation and Universities of the Government of Spain. This study is linked to the project entitled “Descansos activos en la enseñanza superior para la mejora de la atención y rendimiento académico” which is funded under the University of Granada’s 2023 Internal Research Plan (PPJIA2023-124). Data availability The data used to support the findings of the current study are available from the corresponding author upon request. Declarations Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement The data used to support the findings of the current study are available from the corresponding author upon request.