Motor learning research has long established the importance of feedback in refining movement patterns, improving performance accuracy, and supporting skill retention. Traditional feedback forms, such as verbal and visual cues, have been shown to enhance learning efficiency when provided in real time and aligned with the learner’s attentional focus. Previous studies have distinguished between knowledge of performance and knowledge of results feedback, emphasizing that the timing and precision of feedback delivery play critical roles in optimizing motor skill acquisition. The acquisition of motor skills is a complex process that has been extensively studied in the field of sport psychology. Theories of motor learning provide a framework for understanding how athletes learn and refine their movements. Among the most powerful theories, one can distinguish the three-stage model by Fitts and Posner, who suggest the idea that the learners pass three stages: cognitive, associative, and autonomous ones (Kim et al., 2013). The other major theory is the challenge point framework, which holds that an optimum level of learning occurs when the task is complex enough relative to the level of the learner (Buszard, 2022). SDT is a general theory of personality and human motivation (Ntoumanis and Mallett, 2014; Ryan et al., 2019). SDT’s view is that every person possesses three fundamental psychological needs, which are autonomy, competence, and relatedness. With these needs fulfilled, people are more prone to being intrinsically motivated, which results in various positive consequences, among which their well-being, engagement, and performance are to be noted (Ryan and Deci, 2000; Vansteenkiste et al., 2006). Regarding youth sport, SDT offers an insightful model of how to make a positive and encouraging environment for young athletes (Keegan et al., 2009; Balaguer et al., 2013).

Building upon these findings, recent technological advancements have introduced AR as a powerful medium for delivering interactive and adaptive feedback in sports contexts. AR tools allow athletes to visualize performance metrics, motion paths, and spatial targets directly within their environment, promoting immediate error detection and correction. Research in this domain has demonstrated that AR-based interventions can improve precision, movement coordination, and engagement across various sports disciplines. Technology in sports training is not a new thing, since people long ago used video analysis to train, and in the recent past, came the idea of wearable sensors and virtual reality (VR) (Cossich et al., 2023). These technologies have already proven their efficiency in most kinds of sports, including swimming and soccer, and have been applied to the enhancement of such aspects as techniques and tactics, as well as physical condition (Dellaserra et al., 2014; Stone and Kilding, 2009). AR is an enabling technology that superimposes some computer-generated information and images over the real world (Dellaserra et al., 2014). Within sports training, AR may be applied to give the athletes instant feedback about their performance, e.g., displaying the ideal path that a ball should take, or where they are making a technical mistake (Miles et al., 2012). The increasing amount of evidence implies that AR-based feedback may prove an effective means of motor learning and performance (AL-Sinani and Al Taher, 2023; Sattin et al., 2023; Diamond, 2012). The recent innovations in the fields of AR and feedback technologies have raised the level of interest in the sphere of their usage in the development of motor skills, especially among the younger population (Faruk et al., 2025) carried out an experiment that showed the effect of real-time audio-visual feedback in increasing the technical accuracy of novice volleyball players, which in turn results in rapid motor learning. In a similar fashion, (Diller et al., 2025) conducted a mobile AR intervention on elementary school children and found significant differences in gross motor development with the Test of Gross Motor Development-Second Edition (TGMD-2). Li et al. (2025) proposed adaptive hoop training system based on the AR that could dynamically adapt the task difficulty in real-time presented better motor learning results compared to the non-adaptive systems. In a different 2025 work, Diller et al. (2025) introduced SkillAR, a head-worn AR-based system with ongoing spatial feedback that showed success in improving movement patterns with orientation-free corrections. This study followed the methodological principles for controlled sport-science experiments outlined by Roupa et al. (2022), adhered to established wearable-sensor data-collection protocols described in Camomilla et al. (2018), and aligned our reporting procedures with the systematic frameworks for technology-based training interventions recommended by O’Reilly et al. (2018) and Richter et al. (2024).

Although prior research has demonstrated the benefits of both augmented feedback and motivational support for performance learning, limited empirical evidence exists on how real-time AR feedback simultaneously influences motor skill acquisition and intrinsic motivation in team-based youth sports. Moreover, the cognitive mechanisms and motivational pathways underlying these effects remain underexplored. The effect of AR and VR technology on performance and acquisition of different skills in different fields has been investigated in several studies. Couture et al. (1999) proved that AR-based visual feedback enhanced the shooting accuracy and technique significantly in basketball players, and the significance of the real-time visual augmentation. Within the situation of physical education, Gallie and Zhou (2020) discovered that AR-based training resulted in more active student participation and improved skill-building capacities, which favor its educational potential. In the same way, Farley et al. (2020) found that an AR training system during combat sports enhanced the physical fitness and technical performance of the athletes. By generalizing these results to other domains outside sports, Davaris et al. (2019) demonstrated that VR in combination with haptic feedback was effective at improving skill acquisition and retention in surgical training, highlighting the generalizability of immersive technologies to other settings of performance-based learning. A review of the important studies in the field is given in Table 1.

Although the body of literature on using AR in sports training is expanding, some gaps in the literature on the topic exist. To begin with, most of the available literature has concentrated on individual sports, leaving little consideration for group sports. Second, the effect that AR has on motor learning and motivation has not been studied much. Third, research on the application of AR among young athletes is required. The current study will help fill this gap as it examines the effects of a real-time AR feedback system on motor learning and motivation in youth team sports.

A mixed-methods, pre-test/post-test control group design was used to investigate the effects of the AR feedback intervention. Participants were randomly assigned to either the AR feedback group or the control group. The study was conducted over a period of 8 weeks, with a follow-up assessment 4 weeks after the conclusion of the intervention.

Participants were 60 youth soccer players (30 male, 30 female) between the ages of 12 and 15 (M = 13.5, SD = 1.2). All participants had at least 2 years of experience playing organized soccer. Participants were recruited from local soccer clubs and were randomly assigned to the AR feedback group (n = 30) or the control group (n = 30). Random assignment was performed using a stratified randomization procedure to ensure comparable distributions of experience levels between groups. In accordance with the Participant Classification Framework proposed by McKay et al. (2021), all participants were categorized as trained/developmental athletes (Tier 2). This classification reflects individuals engaged in structured, coach-led training programs within organized club environments, typically participating in regional competitions but not yet reaching national or elite levels. The participants in this study trained three to four times per week, received formal technical and tactical coaching, and competed regularly in regional youth leagues. This level of structured engagement and developmental focus aligns with McKay’s criteria for the trained/developmental tier, ensuring appropriate contextualization of the sample within established sport performance standards. Participants were categorized into three experience strata based on their years of organized soccer participation: novice (≤2 years), intermediate (3–4 years), and advanced (≥5 years). Within each stratum, participants were randomly allocated to the AR feedback or control condition in equal proportions. This approach helped maintain balanced representation of varying skill levels and minimized potential bias from pre-existing differences in training background. The mean experience level was 3.2 ± 0.8 years for the AR group and 3.1 ± 0.9 years for the control group, with no significant difference between them [t(58) = 0.41, p = 0.68]. Distribution checks confirmed equivalent proportions of novice, intermediate, and advanced players in both groups, demonstrating that the stratified randomization procedure successfully balanced experience variability. The study involved non-invasive training activities and did not collect or record any personal or sensitive information. In line with institutional policy for non-human-subject research, formal ethical approval was therefore not required. All participants and their parents or guardians were fully informed about the study procedures, and voluntary participation consent was obtained in writing. Table 2 provides a summary of the demographic characteristics of the participants.

The AR feedback intervention consisted of a custom-developed AR application that provided real-time visual feedback on participants’ performance during a series of soccer drills. The AR feedback system used in this study combined real-time visual and auditory cues to assist athletes during training. The visual feedback appeared as digital overlays on the head-mounted display, showing the ball trajectory, passing accuracy, and shot speed in real-world space. The system also produced brief auditory sounds when the players did not follow the target trajectory or performance threshold to indicate the necessity of adjustment. The timing of the feedback was set to match player movement and also had a latency of approximately 150 milliseconds, which was an almost immediate response. The training periods were 45 min, and the participants were given AR feedback during the entire training of all the drills. This integration helped players to monitor the changes in performance and fix movement patterns in real time. The feedback was shown on a head-mounted display (HMD) to the participants and included feedback in terms of ball path, speed of the shot, and accuracy. AR feedback intervention was a personally developed application that was used to offer real-time visual and auditory feedback to a head-mounted display during soccer exercises. The system referred to the use of the Microsoft HoloLens 2 headset with Vicon motion tracking sensors installed around the training area to record the location of the player and the ball. The information was sent to a central processing unit developed in the Unity 3D development package through the AR Foundation SDK of spatial rendering. The latency of the system was an average of one hundred and fifty milliseconds, so that the feedback could be immediate and continuous to the user. Visual overlays were also presented with the provision of ball trajectory lines, an accuracy area in the target, and numeric shot speed indicators. The auditory sound was used to warn about exceeding or falling below preset performance levels. All the feedback elements were supposed to support the correction of the skill and enhance accuracy in execution. The control group would do the same drills as the AR feedback group, but they would be given the regular verbal feedback by a qualified coach.

The motor performance was evaluated by a battery of soccer-specific skill tests, such as a dribbling test, a passing test, and a shooting test. Intrinsic motivation was also measured through IMI (Hendry, 2012), which was a self-report questionnaire that gauged the interest/enjoyment, perceived competence, effort/importance, and perceived choice. To enhance the transparency and applicability in the field, Figure 1 has been given in detail to demonstrate how the AR system was combined with the soccer training environment.

Figure 2 shows the display format, the location of the sensors, as well as the flow of information that every participant would undergo. These additions guarantee clarity to the replication and give a practical insight to the practitioners on how the AR system can be used in a real team training environment. The G Power 3.1 tool was used to justify the sample size by a power analysis. The analysis revealed that at least 52 participants would be enough to identify the medium effects (f = 0.25) with power = 0.80 and α = 0.05. The eventual sample of 60 participants thus satisfied and surpassed this criterion and therefore provided sufficient statistical power to the analyses.

Table 3 summarizes the measurement tools and constructs.

Three subscales of the Intrinsic Motivation Inventory (IMI) measured motivational outcomes: interest, enjoyment, perceived competence, and effort importance. Such dimensions were chosen as they are directly related to the study goals of studying engagement, confidence, and task investment in soccer training. Other IMI subscales, such as pressure, tension, and value usefulness, were not included in order to ease the burden on the participants and to stick to the main points of autonomous motivation that seem to be the most relevant to the feedback-based learning. The choice of these subscales is consistent with the past applied sport motivation research that focuses on self-determination in performing the feedback tasks in the area of performance feedback.

The experiment was carried out within 12 weeks, as shown in Table 4. Both groups of participants underwent two training sessions each week over a period of 8 weeks. The level of motor performance and intrinsic motivation was measured at baseline (Week 1), right after the intervention (Week 8), and a 4-week follow-up (Week 12). Qualitative data were gathered to supplement the quantitative results by use of semi-structured interviews, which were meant to investigate the perceptions of the participants on the feedback experience, motivation, and the learning processes. Ten interviews were carried out (5 in the AR feedback and 5 in the control group) as soon as the training program was finished. The interviews took about 20–25 min, and they were held in a quiet meeting facility within the training facility. The two trained researchers conducted interviews, which were independent of the training sessions, in order to reduce the chances of bias and neutral data collection. Some of the open-ended questions in the interview protocol were: How did the feedback you got influence your motivation during practice? What did you like about the training that helped you acquire new skills, and how did you feel about the timing and the type of feedback you got? All interviews were tape recorded with permission and transcribed word-for-word to be analyzed.

The analysis of the quantitative data was done using SPSS version 28. All variables were calculated using descriptive statistics. To compare the differences between the AR feedback and the control group over time, a two-factor mixed model analysis of variance was conducted, in which one of the factors was between-subjects (group: AR feedback and control) and the other one was within-subjects (time: pre-test, post-test, and follow-up). All the statistical assumptions were verified before conducting the analyses to allow validity. Normality was checked by the Shapiro–Wilk test, the homogeneity of variance was checked using the Levene test, and the Mauchly test was used to test the assumption of being spherical. In the case of a violation of the assumption of sphericity, the Greenhouse–Geisser correction was used to correct the degrees of freedom. All analyses were constructed with a significance threshold of p < 05. In cases where there were any significant group x time interactions, then a pairwise comparison was done with Bonferroni corrections to adjust multiple testing. Besides p-values, effect sizes were also obtained to determine the level of observed differences. The omnibus ANOVA effects were reported with partial eta squared (η²ₚ) values, and the within-group and between-group pairwise comparisons at post-test and follow-up were calculated with Cohen’s d values. All comparisons incorporated 95 % confidence intervals to make both statistical and practical significance interpretations of the results. Lastly, Pearson correlation tests were conducted to determine how changes in motivation scores are related to improvement in motor performance. The correlations were calculated separately, on each group, to determine the potential differences in the way motivational enhancement was related to gains in performance. The principles of the Declaration of Helsinki were followed in the study. No personal or identifying or biometric information was obtained, and all performance information is anonymized before analysis. The participation was limited to non-invasive training exercises done in the standard practice setting of the athletes. The institutional policy states that such non-invasive and fully anonymized research is not to be formally reviewed by the board. Informed consent was received before the inclusion of all participants in the study, both written and verbal consent of the parents and legal guardians of all the participants, respectively.

Although the results of the research conducted in this paper offer valuable information on the importance of AR feedback in the sport training of youth, a number of limitations must be acknowledged. First, the intervention time was relatively short, and it was not possible to determine long-term retention and consolidation of motor learning. Longitudinal designs should also be used in future research to establish whether the recorded performance and motivation changes are maintained in the long term and at various phases of athletic growth. Second, the sample was a group of youth athletes that have been trained in one regional academy only, which could limit the application of the results to other sporting conditions or levels of competition. External validity would be enhanced by increasing the research in various institutions, different kinds of sports, and different ages. Third, the research was aimed at behavioral and self-report outcomes, and it was not concerned with direct measurement of underlying cognitive and neural processes. Future research may involve motion capture, eye-tracking, or neurophysiological research to learn more about how AR feedback affects the state of attentional focus, motor planning, and feedback processing. Lastly, although the motivational results agreed with the postulates of SDT, the current design failed to measure autonomy, competence, or relatedness as mediation variables. These mechanisms should be empirically tested in future research by using validated SDT-based need satisfaction scales. Overcoming these obstacles, future research will help to learn more about the role of AR technology in promoting motor learning and motivation, which will eventually allow for introducing more effective and evidence-based changes in the context of the implementation of digital tools into the youth sport training setting.