Effects of virtual reality-based cognitive and technical drills on scanning and passing performance in youth football players: a randomized controlled study Authors: Kaan Erişik, Ali Onur Cerrah, Fahri Safa Cinarli Published: 2025 ABSTRACT AbstractBackgroundTraditional training methods often fall short in replicating the perceptual load of match environments. Virtual reality (VR) has emerged as a promising modality to enhance cognitive-motor integration in football contexts. Purpose: This study aimed to examine the effects of a 7-week virtual reality-based training program designed to improve both scanning behavior and passing performance in youth football players, in comparison with traditional training methods. Methods: A parallel-group randomized controlled trial was conducted with 22 male youth players from U16–U17 squads (mean age = 16.77 ± 0.42 years), who were assigned to either a VR group (n = 11) or a control group (n = 11). The VR group completed 3 weekly sessions using the SensiballVR™ platform for a duration of 7 weeks, in addition to their regular training. Scanning frequency (before ball reception, during control, and off-the-ball) and passing performance (by execution type, outcome, and pass style) were assessed via video analysis in small-sided games pre- and post-intervention. Results: Between-group analysis revealed that the VR group achieved significantly greater improvements across all scanning domains compared to the control group, with mean-based percentage increases (calculated as the average of individual relative changes) ranging from + 198% to + 456%, rising from group mean values of 18 to 43 scans before ball reception, 20 to 43 during ball control, and 257 to 714 off the ball, versus − 12% to + 37% in controls (p < 0.01; ES = 0.58–0.83). Within-group analysis confirmed that scanning frequency increased significantly from pre- to post-test in the VR group (p = 0.003, ES = 0.89), whereas only off-the-ball scanning improved in the control group (p = 0.008, ES = 0.81). In passing performance, the VR group improved significantly in one-touch short-successful passes (+ 38%, p = 0.006, ES = 0.83), whereas the control group showed no meaningful change (+ 28%, p = 0.247, ES = 0.35). Although the VR group initially showed higher performance in control-pass short-successful passes (p = 0.038; ES = 0.44), this difference did not remain significant after false discovery rate correction (q = 0.199). Under the same condition, penetrative passes also improved significantly within the VR group (+ 108%, p = 0.029, ES = 0.66). No significant effects were observed for multi-touch passes (p > 0.05). Conclusion: Preliminary evidence suggests that immersive VR-based training can meaningfully enhance scanning behavior in youth football players, while potential benefits for passing outcomes remain exploratory. These findings highlight VR technology as a promising complementary tool in modern football development frameworks for improving perceptual-cognitive skills and decision-making.Clinical trial numberClinicalTrials.gov identifier: NCT07144371 (retrospectively registered on 27/08/2025).Keywords: Virtual reality, Cognitive training, Scanning behavior, Passing performance, Perceptual skills Introduction Football is a multifaceted sport that demands not only physical prowess but also advanced cognitive and perceptual skills, particularly in fast-paced and decision-heavy game situations [1, 2]. Among these, environmental scanning has been recognized as a critical component influencing both individual and collective performance [3]. Effective scanning enhances players’ ability to anticipate game developments, select appropriate passing options, and maintain tactical awareness under pressure [4]. Passing performance is a core technical skill that relies heavily on the integration of perceptual input and rapid motor execution. Meta-analytic evidence on perceptual–cognitive expertise in sport further indicates that expert athletes employ more efficient visual search strategies and achieve higher response accuracy, capacities that likely underpin behaviors such as frequent scanning and more contextually appropriate passing decisions observed in high-level youth and professional players [5]. Despite the known importance of these intertwined skills, traditional training methods often fall short of replicating the cognitive load and complexity of in-game scenarios, particularly for developing athletes [6]. In recent years, virtual reality (VR) has emerged as a promising tool to bridge this gap. By providing immersive, interactive, and customizable environments, VR-based interventions offer a safe and controlled platform for simulating complex game scenarios [7]. From a neurophysiological perspective, VR training can stimulate the activation of key brain regions involved in perception, attention, and decision-making [8]. Recent research examined the relationship between perceptual-cognitive performance in a dynamic 360° environment and soccer-specific performance in youth players [9]. Their findings indicated a significant positive correlation between performance in a 360° multiple object tracking task and both passing accuracy and defensive effectiveness in small-sided games. The repeated exposure to dynamic, visually rich stimuli enhances selective attention, working memory, and anticipatory skills, which are all critical for effective environmental scanning [10, 11]. Furthermore, the real-time feedback and motor engagement involved in VR-based drills promote sensorimotor integration, strengthening the neural pathways that support technical execution, such as passing accuracy and timing [12]. These findings suggest that VR is not merely a visual aid but a cognitive training modality capable of eliciting measurable changes in sport-specific decision-making and technical performance [13]. However, current literature is still limited in terms of randomized controlled trials investigating the combined effects of VR on scanning and passing performance in football players. Therefore, this study aimed to evaluate the effects of a VR-based training program that integrates cognitive and technical drills on scanning behavior and passing performance in youth footballers. It was hypothesized that VR-based training intervention would demonstrate significantly greater improvements in scanning frequency and passing accuracy compared to those in the control group receiving only traditional training. Methods Study design and participants This parallel-group randomized controlled trial aimed to investigate the effects of a virtual reality–based training program on scanning behavior and passing performance in youth football players. The study was conducted during the pre-season period in collaboration with a regional football academy. A total of 22 male youth players from the U16 and U17 age categories volunteered to participate. This study was conducted in accordance with the CONSORT 2025 guidelines for reporting randomized controlled trials. All participants and their legal guardians provided informed consent prior to participation. The study was approved by the Ethics Committee of Eskişehir Technical University (Approval date and number: 17/03/2023, E-87914409-050.03.03.03.04–2300016243) and was conducted in accordance with the principles of the Declaration of Helsinki. Inclusion criteria required players to be injury-free in the past 12 months, possess normal or corrected-to-normal vision, and have no prior exposure to VR-based training. Players with any neurological or vestibular disorders or recent orthopedic surgery were excluded. As no directly comparable trials of VR-based scanning or passing interventions were available in the literature, a priori sample size estimation was not feasible; therefore, power was assessed post hoc after study completion. A post-hoc power analysis was performed to evaluate the adequacy of this sample size (n = 11 per group). Effect sizes derived from nonparametric test statistics were converted to Cohen’s d for approximate estimation. Power was > 0.80 for large effects (𝑟 = 0.78–0.83), moderate for medium effects (r = 0.44), and low-to-moderate for smaller effects (r = 0.26–0.36), indicating sufficient sensitivity to detect large but not small-to-moderate effects. Stratified randomization was performed based on age category (U16 vs. U17) and playing position, using a computer-generated random sequence. The allocation procedure was conducted by a researcher not involved in coding or outcome assessment. While the control group continued regular field-based training only, the VR group received additional virtual reality training sessions three times per week for 7 consecutive weeks using the SensiballVR™ platform. The drills were progressively intensified in terms of visual load and technical complexity. Body composition measurement Anthropometric assessments were conducted to characterize the physical profiles of the participants. Standing height was measured to the nearest 0.1 cm using a wall-mounted stadiometer (Holtain Ltd., UK), with participants standing barefoot in an upright anatomical position. The device was calibrated and level-checked prior to each measurement session to ensure accuracy. Body mass was recorded using a digital scale (Tanita BF-522, Japan) with a sensitivity of ± 0.1 kg, while participants wore light training clothing and no shoes. All measurements were taken in the morning hours to minimize the effects of daily biological fluctuations. Small-Sided game protocol Environmental scanning and passing performance were evaluated across four structured small-sided game sessions, comprising two formats: a 5v5 + 1 configuration (4 × 5 min bouts, 2-min rest) and a 7v7 + 7 joker format (3 × 7 min bouts, 2-min rest). Each format was performed twice on separate days within the same week. These formats were chosen to increase the frequency of ball contact, decision-making opportunities, and perceptual-motor demands in a controlled setting. For statistical analysis, player data from the four sessions were averaged to yield a single pre- and post-test value. All games were conducted on synthetic turf under consistent environmental and temporal conditions. Players were instructed to compete at maximum intensity, and verbal encouragement was used to sustain engagement throughout. All sessions were recorded using synchronized aerial and ground-level video systems, including DJI Mini 3 and Air 2S drones (DJI, China) and iPhone 13 cameras (Apple Inc., USA), to capture comprehensive positional and behavioral data. These recordings served as the source for both scanning and passing analyses. Coding was conducted independently by two analysts trained in notational analysis. Coders were blinded to group allocation. Prior to the main analyses, both analysts participated in a pilot coding session using representative sequences from small-sided games. During this session, the operational definitions of scanning and passing categories were reviewed and trial runs were jointly discussed until full agreement was achieved. This preparatory training ensured consistent application of the coding scheme. To ensure accuracy and consistency, the videos were reviewed in slow motion and frame-by-frame where necessary. Inter-rater reliability was assessed using a two-way mixed-effects, absolute agreement model [ICC (3, 1)], yielding excellent agreement (ICC = 0.92, 95% CI [0.86–0.97]).” Scanning performance Each scanning event was classified into one of three predefined categories: before ball reception, during control, and off-the-ball. These classifications were based on distinct head movement patterns and intentional visual searches, as detailed in Table 1. Table 1Operational definitions of scanning categoriesTermDefinitionBefore Ball ReceptionHead movements and visual search actions performed before receiving the ball, typically during off-the-ball movement.During ControlBrief glances away from the ball to assess the environment while maintaining control.Off-the-BallScanning actions not directly related to immediate ball reception, including awareness checks and positional tracking. Operational definitions of scanning categories Passing performance Each pass was categorized based on three main criteria: the type of execution, the outcome, and the nature of the pass. In terms of execution, passes were classified as one-touch, controlled, or those made after three or more touches. The outcome dimension referred to whether the pass was successful-i.e., completed to a teammate-or unsuccessful, such as when intercepted or misplaced. Finally, passes were further classified by type, including short, deep (penetrative), and long (aerial) deliveries. All pass types were coded according to the operational definitions presented in Table 2. Table 2Operational definitions of passing categories used in the analysisTermDefinitionOne-touch PassA pass executed immediately after receiving the ball, without a second touchControlled PassA pass made after a second touch but before the third, following initial ball receptionPass with ≥ 3 TouchesA pass executed after making three or more touches on the ball before delivering itSuccessful passA pass that successfully reaches a teammateUnsuccessful passA pass that reaches an opponent or leads to loss of possessionShort passAny ground pass that does not meet the criteria for a deep passPenetrative passA ground pass that eliminates at least two opposing players before reaching a teammateLong passAny pass delivered through the air to another player Operational definitions of passing categories used in the analysis Virtual reality training system The intervention consisted of a 7‑week virtual reality-based training program designed to enhance scanning behavior and passing performance in youth football players. The virtual reality training intervention in this study was delivered through the SensiballVR™ system, a patented immersive training technology developed for football-specific cognitive and technical development. The system operates via a high-performance desktop computer connected to HTC Vive Pro Eye headsets (HTC Corporation, Taiwan) and integrates a pair of custom-designed haptic feedback shoes that allow for realistic foot-to-ball interaction. These haptic shoes (SensiballVR™ BLE) provided vibration-based tactile feedback with a 7.5 G amplitude, delivered through Bluetooth 5 connectivity and controlled via the SensiballVR software. The vibration cues were synchronized with ball contact and ground interactions to simulate authentic tactile sensations and enhance sensorimotor realism during the virtual drills. SensiballVR™ incorporates a proprietary sensor system and training software that enables real-time tracking of foot movements and precise environmental interaction within a 4 × 4 m training space. Unique to this setup is the patented haptic footwear system, which works in conjunction with HTC Vive trackers to provide both tactile feedback and high-resolution motion capture. Although no prior validation studies exist specifically for SensiballVR™, recent systematic reviews indicate that immersive simulation technologies, including head-mounted displays with motion tracking, can effectively replicate perceptual–cognitive and technical demands in football training contexts [14]. In addition, feasibility analyses of VR applications in football confirm that such systems can simulate realistic scenarios, overcome environmental limitations, and support the acquisition of technical and perceptual skills [15]. The training software includes a variety of progressively challenging scenarios targeting multi-object tracking, shoulder-scan tasks, peripheral vision, and time-pressured passing decisions to simulate real-game cognitive demands. Representative examples of these VR training drills are illustrated in Fig. 1. Scenarios were designed to maintain ecological validity by integrating realistic teammate and opponent movements as well as game-relevant decision-making cues. All performance data, including scanning frequency and passing accuracy, were automatically recorded through the integrated reporting tools of the SensiballVR™ platform. Composite illustration of VR-based football training drills, including working-memory tasks (A), peripheral-vision challenges (B), shoulder-scan drills (C), and multi-object tracking exercises (D) Intervention protocol Training sessions were held three times per week over a 7‑week period (total of 21 sessions), and were implemented in addition to the participants’ routine football training. The VR drills were specifically structured to target two core domains: (1) environmental scanning, through tasks requiring frequent and purposeful visual exploration of the surrounding virtual pitch before decision execution, and (2) technical passing, via simulated teammate and opponent interactions demanding timely and accurate ball delivery. All drills were performed using a sensor-integrated VR ball to maintain ecological validity and ensure motor engagement. Each session was supervised by a trained operator to ensure proper technique, consistency, and safety. To support motor-cognitive adaptation and avoid learning plateaus, the training protocol was structured into three progressive phases over the 7‑week period, with each phase focusing on increasingly complex visual and technical demands (Table 3). The control group, in contrast, continued with regular team practices and did not receive any VR-based training during the intervention period. Table 3Progressive structure of the SensiballVR™ training protocolWeeksTraining Focus1–2Familiarization and basic-level visual exploration drills with limited distractors.3–5Intermediate tasks with increased visual stimuli, added decision-making components, and time constraints.6–7Advanced drills involving dual-tasking, simulated pressure, and complex passing under cognitive load. Progressive structure of the SensiballVR™ training protocol Statistical analysis All statistical analyses were conducted using IBM SPSS Statistics (Version 23, NY, USA). Due to the relatively small sample size and the nature of the performance-based outcome measures, the data were first evaluated for normality using the Shapiro–Wilk test. Baseline comparisons of demographic and performance variables between groups were conducted using independent-samples t-tests or Mann–Whitney U tests, depending on distributional assumptions, to confirm group equivalence at pre-test. As several variables violated the assumption of normal distribution, non-parametric tests were employed throughout the analysis. Within-group comparisons were carried out using the Wilcoxon signed-rank test to assess changes from pre- to post-intervention in both the VR intervention group and the control group. To evaluate between-group differences, the Mann–Whitney U test was applied to individual delta scores (post–pre). Because individual baseline variability was expected to be high, both median delta values and mean percentage changes were computed. Median delta values represent the median of individual raw differences, while percentage changes were calculated as the average of individual relative changes [(meanpost − meanpre)/meanpre] × 100 to illustrate the overall magnitude of improvement across participants. While mean-based percentage change provides a general central trend, median-based delta offers a robust measure against non-normality. This dual reporting approach was intended to ensure both statistical rigor and practical interpretability of the results. To control for the inflation of Type I error due to multiple between-group comparisons, the Benjamini–Hochberg false discovery rate (FDR) correction was applied across all scanning and passing outcomes (m = 21). Effects with q < 0.05 (FDR-adjusted p values) were considered statistically significant, whereas those with higher q values were interpreted as exploratory. Effect sizes were calculated using the formula r = Z/√N. These values were interpreted in accordance with conventional thresholds, with r values of 0.10, 0.30, and 0.50 indicating small, moderate, and large effects, respectively. The significance level for all tests was set at p < 0.05. Results Demographic variables were assessed, and the analysis indicated no significant differences between groups at baseline for age (p = 0.143), height (p = 0.560), weight (p = 0.655), or body mass index (p = 0.838) (Table 4). Table 4Demographic variables of football players*GroupsAge (years)Height (cm)Weight (kg)BMI (kg/m2)VR (n = 11)16.90 ± 0.30177.72 ± 6.8864.54 ± 7.4220.38 ± 1.20CG (n = 11)16.63 ± 0.50179.09 ± 3.3065.72 ± 4.4220.48 ± 1.12p value0.1430.5600.6550.838*Data are presented in mean ± standard deviation. VR Virtual reality group, CG Control group, BMI Body mass index. p values represent the value of Independent sample T test. Demographic variables of football players* *Data are presented in mean ± standard deviation. VR Virtual reality group, CG Control group, BMI Body mass index. p values represent the value of Independent sample T test. Baseline performance comparisons showed no statistically significant differences between the VR and control groups in any scanning or passing measures (all p > 0.05) (Table 5). Table 5Baseline comparisons of scanning and passing measures between groups (pre-test values)*VariableVR (n = 11)CG (n = 11)Test statisticp valueBefore ball reception18.00 ± 13.1124.90 ± 16.29−1.096 a0.286During control20.18 ± 13.0419.45 ± 18.20−0.108 a0.915Off-the-ball257.00 ± 206.28213.90 ± 125.410.592 a0.560One touch short-successful46.45 ± 12.9542.18 ± 18.240.633 a0.534One touch short-unsuccessful10.00 ± 4.819.90 ± 4.740.045 a0.965One touch penetrative-successful5 (2–8)6 (2–10)−0.699 b0.484One touch penetrative-unsuccessful1 (1–3)2 (1–4)−0.169 b0.866One touch long-successful1 (0–4)1 (0–2)−1.396 b0.163One touch long-unsuccessful1 (0–3)0 (0–1)−1.237 b0.216Control pass touch short-successful38.27 ± 9.0436.81 ± 12.790.308 a0.761Control pass short-unsuccessful7.18 ± 4.836.72 ± 2.720.272 a0.789Control pass penetrative-successful8.72 ± 4.408.72 ± 5.380.000 a1.000Control pass penetrative-unsuccessful1 (0–2)2 (1–3)−1.882 b0.060Control pass long-successful3.00 ± 2.601.81 ± 1.611.281 a0.218Control pass long-unsuccessful0 (0–0)1 (0–1)−1.535 b0.125Multi touch short-successful11.09 ± 4.329.18 ± 6.250.832 a0.415Multi touch short-unsuccessful1 (0–4)2 (1–4)−0.703 b0.482Multi touch pass penetrative-successful2.72 ± 1.902.09 ± 1.810.803 a0.431Multi touch penetrative-unsuccessful0 (0–1)1 (0–2)−0.795 b0.426Multi touch long-successful1 (0–2)0 (0–1)−1.030 b0.303Multi touch long-unsuccessful0 (0–0)0 (0–1)−1.481 b0.243*Values are presented as Mean ± SD for normally distributed variables and as Median (25th–75th percentile) for non-normally distributed variables. VR Virtual reality group; CG Control group. ᵃ Independent-samples t-test, ᵇ Mann–Whitney U test. Baseline comparisons of scanning and passing measures between groups (pre-test values)* *Values are presented as Mean ± SD for normally distributed variables and as Median (25th–75th percentile) for non-normally distributed variables. VR Virtual reality group; CG Control group. ᵃ Independent-samples t-test, ᵇ Mann–Whitney U test. Between-group comparisons based on delta values revealed that the VR group significantly outperformed the control group across all scanning variables (p < 0.01). These differences remained significant after false discovery rate correction (q ≤ 0.042), with large effect sizes for scanning before ball reception (ES = 0.83), during ball control (ES = 0.58), and off-the-ball actions (ES = 0.78). These results highlight the robustness, practical relevance, and effectiveness of the VR-based training intervention (Table 6). Table 6Changes in scanning performance in VR and control groups (n = 22)*Scanning variablesGroupMedianPre(P25-P75)MedianPost(P25-P75)Median Delta p qESBefore ball receptionVR17 (7–28)37 (21–57)+ 22< 0.0010.0100.83CON23 (13–36)23 (8–34)−1During controlVR25 (4–33)39 (30–52)+ 210.0060.0420.58CON13 (4–36)18 (4–40)+ 2Off-the- ballVR190 (94–458)389 (354–1640)+ 237< 0.0010.0100.78CON199 (154–361)259 (189–375)+ 54*Data are presented in median and interquartile range (P25–P75). p values represent between-group comparisons of delta scores using the Mann–Whitney U Test; q values represent false discovery rate (Benjamini–Hochberg) adjusted p values; ES Effect size Changes in scanning performance in VR and control groups (n = 22)* *Data are presented in median and interquartile range (P25–P75). p values represent between-group comparisons of delta scores using the Mann–Whitney U Test; q values represent false discovery rate (Benjamini–Hochberg) adjusted p values; ES Effect size Within-group analyses showed significant improvements in the VR group across all scanning domains: before ball reception (p = 0.003, ES = 0.89), during control (p = 0.003, ES = 0.89), and off-the-ball (p = 0.003, ES = 0.89). In contrast, the control group showed a significant gain only in off-the-ball scanning (p = 0.008, ES = 0.81), with no meaningful changes observed in the other domains (Fig. 2). Changes in scanning frequency before ball reception, during ball control, and off-the-ball across time (T1–T2) in VR and control groups. The y-axis represents the number of scanning actions per player per game (ppg). Data are presented as median and interquartile range The VR group demonstrated marked improvements in all scanning domains, with mean percentage increases of + 456% before ball reception (corresponding approximately to an increase from group mean values of 18 to 43 scans per game), + 223% during ball control (from 20 to 43), and + 199% off the ball (from 257 to 714). These percentages were calculated as the average of individual relative (within-subject) changes, while the absolute values represent group means provided for contextual interpretation. In contrast, the control group showed only minimal changes (−12% to + 37%) across the same domains (Fig. 3). It should be noted that Fig. 3 depicts mean-based percentage changes, which illustrate the overall magnitude of improvement across participants, whereas Table 6 reports median delta values representing the median of individual changes. Mean percentage change in scanning frequency across three phases of play (before the ball reception, during ball control, and off-the-ball) in the VR and control groups Although between-group comparisons did not reach statistical significance (all p > 0.05; FDR-adjusted q > 0.05), all successful one-touch pass types showed positive trends in favor of the VR group. Small-to-moderate effect sizes (ES = 0.16–0.32) suggest modest yet consistent improvements following the intervention, which should be interpreted as exploratory (Table 7). Table 7Changes in passing performance following one touch in the VR and control groups (n = 22)*Pass TypeGroupsMedianPre(P25-P75)MedianPost(P25-P75)Median Delta p qESShort-successfulVR48 (36–54)66 (45–77)+ 140.2230.4700.26CON46 (19–54)46 (38–70)+ 10Short-unsuccessfulVR9 (6–13)8 (5–11)−30.2100.4700.27CON12 (7–12)11 (6–14)+ 1Penetrative-successfulVR5 (2–8)8 (6–13)+ 30.4470.6690.16CON6 (2–10)8 (2–10)+ 1Penetrative-unsuccessfulVR1 (1–3)1 (0–3)00.6650.7610.09CON2 (1–4)1 (0–2)0Long-successfulVR1 (0–4)2 (0–2)−10.1340.4690.32CON1 (0–2)1 (0–2)0Long-unsuccessfulVR1 (0–3)0 (0–1)00.2650.5050.24CON0 (0–1)0 (0–1)0*Data are presented in median and interquartile range (P25–P75). p values represent between-group comparisons of delta scores using the Mann–Whitney U Test; q values represent false discovery rate (Benjamini–Hochberg) adjusted p values; ES = effect size Changes in passing performance following one touch in the VR and control groups (n = 22)* *Data are presented in median and interquartile range (P25–P75). p values represent between-group comparisons of delta scores using the Mann–Whitney U Test; q values represent false discovery rate (Benjamini–Hochberg) adjusted p values; ES = effect size Within-group analysis showed that the VR group demonstrated a statistically significant improvement in one-touch short-successful passes from pre- to post-test (p = 0.006, ES = 0.83), while the control group also showed a non-significant increase in the same metric (p = 0.247, ES = 0.35). Although moderate-to-large effect sizes were observed in other one-touch pass types within the VR group (e.g., penetrative-successful: ES = 0.51; long-successful: ES = 0.33), these changes did not reach statistical significance (p > 0.05). The control group exhibited trivial to small effects across all one-touch passing metrics (ES = 0.10–0.30), indicating a limited response to the intervention (Fig. 4). Changes in one-touch passing performance (successful and unsuccessful) across different pass types (short, penetrative, and long) from T1 to T2 in the VR and control groups. The y-axis represents the number of passes per player per game (ppg). Data are presented as median and interquartile range Between-group comparisons initially indicated a statistically significant improvement in short-successful passes during the control-pass condition in the VR group (p = 0.038, ES = 0.44), while the control group exhibited a decrease in performance (Table 8). However, this effect did not remain significant after false discovery rate correction (q = 0.199), and should therefore be interpreted as exploratory. Moderate effect sizes in penetrative-successful (ES = 0.30) and penetrative-unsuccessful (ES = 0.36) passes in the VR group during the same condition further suggest a favorable but non-significant trend (p > 0.05), whereas no meaningful differences were observed in the remaining pass categories. Table 8Changes in passing performance following control-pass in the VR and control groups (n = 22)*Pass TypeGroupsMedianPre(P25-P75)MedianPost(P25-P75)Median Delta p qESShort-successfulVR37 (31–48)39 (37–50)+ 50.0380.1990.44CON37 (25–45)28 (22–41)−5Short-unsuccessfulVR7 (3–9)6 (3–8)+ 10.6440.7610.10CON7 (5–9)7 (4–10)0Penetrative-successfulVR9 (5–13)12 (9–16)+ 50.1660.4700.30CON10 (3–14)6 (4–14)−1Penetrative-unsuccessfulVR1 (0–2)1(1–3)+ 10.0930.3900.36CON2 (1–3)1 (0–3)−1Long-successfulVR2 (1–5)3(1–4)00.3180.5360.21CON2 (0–3)2 (1–5)0Long-unsuccessfulVR0 (0–0)0 (0–0)00.7250.7610.08CON1 (0–1)0 (0–2)0*Data are presented in median and interquartile range (P25–P75). p values represent between-group comparisons of delta scores using the Mann–Whitney U Test; q values represent false discovery rate (Benjamini–Hochberg) adjusted p values; ES = effect size Changes in passing performance following control-pass in the VR and control groups (n = 22)* *Data are presented in median and interquartile range (P25–P75). p values represent between-group comparisons of delta scores using the Mann–Whitney U Test; q values represent false discovery rate (Benjamini–Hochberg) adjusted p values; ES = effect size In the VR group, within-group analysis showed a significant improvement in penetrative-successful passes during the control-pass condition (p = 0.029, ES = 0.66), indicating a large effect. Short-successful passes also showed a positive but non-significant increase (p = 0.155, ES = 0.43). Small improvements were observed in penetrative-unsuccessful and long-unsuccessful passes, although these did not reach statistical significance. Conversely, the control group displayed a slight, non-significant decline in short-successful passes (p = 0.109, ES = 0.48) and small-to-moderate effects across other pass types, none of which were statistically meaningful (Fig. 5). Changes in control passing performance (successful and unsuccessful) across different pass types (short, penetrative, and long) from T1 to T2 in the VR and control groups. The y-axis represents the number of passes per player per game (ppg). Data are presented as median and interquartile range No statistically significant changes were observed in any multi-touch pass type between groups (p > 0.05; FDR-adjusted q > 0.05). Median differences were minimal, and all effect sizes remained small (ES ≤ 0.26), indicating limited impact of the intervention on these passing actions (Table 9). Table 9Changes in passing performance following multi-touch in VR and control groups (n = 22)*Pass TypeGroupsMedianPre(P25-P75)MedianPost(P25-P75)Median Delta p qESShort-successfulVR12 (8–13)8 (6–13)+ 10.2240.4700.26CON7 (4–13)11 (6–14)+ 1Short-unsuccessfulVR1 (0–4)2 (0–4)00.7970.7970.06CON2 (1–4)4 (1–6)+ 1Penetrative-successfulVR3 (1–4)3 (2–6)00.4780.6690.16CON2 (0–4)1 (0–3)0Penetrative-unsuccessfulVR0 (0–1)0 (0–0)00.5190.6810.15CON1 (0–2)0 (0–1)0Long-successfulVR1 (0–2)0 (0–1)00.6990.7610.09CON0 (0–1)0 (0–1)0Long-unsuccessfulVR0 (0–0)0 (0–0)00.3320.5360.25CON0 (0–1)0 (0–1)0*Data are presented in median and interquartile range (P25–P75). p values represent between-group comparisons of delta scores using the Mann–Whitney U Test; q values represent false discovery rate (Benjamini–Hochberg) adjusted p values; ES = effect size Changes in passing performance following multi-touch in VR and control groups (n = 22)* *Data are presented in median and interquartile range (P25–P75). p values represent between-group comparisons of delta scores using the Mann–Whitney U Test; q values represent false discovery rate (Benjamini–Hochberg) adjusted p values; ES = effect size Within-group analyses showed no statistically significant changes in any multi-touch pass category for either the VR or control groups (p > 0.05). Effect sizes were generally negligible to small (ES < 0.30), with some reaching the moderate range (ES = 0.31–0.38), suggesting only a minimal impact of the intervention on these passing actions (Fig. 6). Changes in multi-touch passing performance (successful and unsuccessful) across short, penetrative, and long pass types from T1 to T2 in the VR and control groups. The y-axis represents the number of passes per player per game (ppg). Data are presented as median and interquartile range Figure 7 illustrates the percentage changes in successful and unsuccessful passes across one-touch, control-pass, and multi-touch executions. In the VR group, the largest gains were evident in one-touch penetrative passes (+ 157%) and multi-touch short passes (+ 105%), accompanied by notable reductions in unsuccessful passes. Control-pass improvements were more modest, with penetrative passes increasing by + 108%. In contrast, the control group demonstrated minimal or negative changes across most pass categories, particularly in long-pass executions, where decreases of up to − 80% were observed. Mean percentage change in successful and unsuccessful passes during one-touch, control-pass, and multi-touch actions across different pass types (short, penetrative, long) in the VR and control groups Discussion This study investigated the impact of an immersive VR program integrating cognitive-perceptual and technical components on scanning behavior and passing performance in youth football players. The results provide compelling evidence that VR-based training can elicit substantial improvements in visual exploratory actions and passing efficacy, particularly under conditions requiring rapid decision-making and one-touch execution. Notably, players in the VR group demonstrated marked increases in scanning frequency across all phases of ball interaction. Similarly, meaningful enhancements were observed in successful penetrative and short passing actions while unsuccessful pass frequencies either declined or remained stable. Collectively, these findings support the integration of immersive VR technologies as a complementary tool in youth football training to bridge perceptual-cognitive skills with technical execution on the field. Small-sided games have long been recognized as a cornerstone of football training because they reproduce the technical, tactical, and physiological demands characteristic of competitive match play. Previous research has shown that small-sided games can elicit heart rate, blood lactate, and perceived exertion responses comparable to, or even higher than, those recorded during official matches [16, 17]. Such outcomes are strongly influenced by contextual variables, including the number of players involved, pitch size, and specific rule modifications [18]. These findings collectively highlight the ecological validity of small-sided games as representative training environments that maintain the perceptual-cognitive and physical dynamics inherent to real matches. Within this framework, the present study examined changes in scanning and passing performance during small-sided games rather than under full 11-a-side match conditions. Consequently, the observed improvements should be interpreted as enhancements in small-sided game performance rather than as direct evidence of improvements in competitive match outcomes. Nevertheless, because small-sided games closely mirror the physiological and tactical requirements of official match play, it is plausible that the perceptual–technical adaptations observed here may partially transfer to competitive environments. Future studies employing full-match analyses and in-game behavioral metrics are warranted to determine the extent and persistence of this transfer. At the team level, passing performance is widely recognized as a decisive determinant of match success. Recent elite tournament analyses provide important benchmarks for interpreting the present findings. For example, UEFA Euro 2024 analyses showed that winning teams exhibited more backward passes and higher overall passing accuracy/possession, underscoring the importance of controlled circulation for phase construction and tempo management [19]. Similarly, analyses of the FIFA Club World Cup 2025 indicated that passing volume and completion strongly predicted match outcomes, with approximately 478 completed passes distinguishing winners and an ≈ 9% increase in win probability for every additional 100 completed passes [20]. In this context, the present improvements in one-touch short-successful and penetrative passes observed in the VR group are not isolated technical gains but represent mechanisms that enable youth players to move closer to these elite-level thresholds of passing efficiency and possession stability. Enhanced scanning behavior likely facilitated faster option generation, thereby increasing one-touch success and contributing to the overall passing volume and accuracy that are critical for success at higher competitive levels. The most pronounced enhancements identified in this investigation pertained to scanning frequency across all three temporal phases of gameplay (r >0.89). While previous research has emphasized the critical role of scanning behavior in football and demonstrated its sensitivity to task constraints, such as increased scanning rates during small-sided formats like 4v4 [21], no prior study has directly investigated the effects of immersive VR training on scanning frequency itself. In this context, existing evidence suggests that VR-based training can indirectly enhance perceptual-cognitive skills, including visual exploration and tactical awareness [22], which are fundamental to effective scanning in dynamic football contexts [3]. Similarly, in officiating domains, VR environments have been shown to replicate on-field scanning patterns with high fidelity; for example, sub-elite football referees demonstrated a 45% overlap between VR and live match gaze behaviors, compared to only 35% in traditional video conditions [23]. This finding reinforces the representativeness of VR scenarios and supports their potential to cultivate realistic visual exploration strategies in both players and officials. To our knowledge, this is the first study to systematically analyze scanning frequency across three distinct temporal phases of gameplay within a virtual reality training context. This temporal sensitivity in scanning behavior underscores the functional relevance of the observed improvements. These enhancements likely stem from the combined effects of perceptual-cognitive adaptation and VR-induced attentional engagement. Exposure to dynamic and time-pressured scenarios appears to sharpen selective attention, working memory, and anticipatory planning [24]. Neurocognitively, these processes may engage the frontoparietal attention network, particularly the dorsolateral prefrontal cortex and posterior parietal cortex, which are central to spatial orientation and decision-making under pressure [25]. Beyond these neurocognitive mechanisms, it is also worth considering whether scanning behavior reflects inherent creativity or primarily adapts to situational task demands. Interestingly, both highly creative and less creative players demonstrated similar gaze strategies during football passing tasks, indicating that visual exploration is governed more by contextual constraints than by individual creative ability [26]. This perspective suggests that improvements in scanning are attainable through structured and repeated exposure to representative scenarios rather than being an exclusive trait of inherently creative players. The observed enhancements in passing performance provide additional empirical support for the integrative potential of VR-based interventions in football training. Statistically significant improvements emerged in one-touch short-successful passes and control-based short passes within the VR group, with percentage gains in successful penetrative and short passes reaching + 157% and + 108%, respectively. These results indicate that perceptual-cognitive adaptations fostered by immersive VR training may translate into more accurate and efficient motor execution under time-constrained, high-pressure scenarios typical of rapid one-touch play. Supporting this, a 12‑week randomized controlled trial reported a 53.1% improvement in passing performance and a 25.3% enhancement in shooting accuracy among junior football players following VR‑supported training, whereas the control group demonstrated negligible gains [27]. Similarly, in an 8‑week randomized trial with 26 youth soccer athletes (VR, n = 13; video, n = 13), both VR and video groups improved significantly in passing decision‑making (p < 0.01) and visual search behavior (p < 0.01), yet the VR group achieved markedly greater gains, as evidenced by a significant group × time interaction (p < 0.01) [8]. The concurrent improvements observed in scanning frequency and passing accuracy in the VR group are likely mechanistically linked. Enhanced visual exploration offers richer contextual information regarding teammate positioning, opponent pressure, and available passing lanes, thereby facilitating more adaptive and precise motor execution. This relationship is supported by evidence showing that players who scan more frequently prior to receiving the ball execute faster and more accurate passes under pressure [28, 29]. A controlled study involving U13 and U23 elite players further demonstrated that faster passing performance was significantly associated with more head turns before ball reception and fewer during possession, particularly in older players (p < 0.01) [30]. Moreover, VR-based interventions further support perception–action integration by recreating time-pressured scenarios that simultaneously train attentional control and technical execution [22]. Collectively, these results reinforce the importance of prospective visual exploration as a precursor to efficient passing execution. In this view, the perception–action coupling fostered by VR may partly explain the substantial improvements in passing efficiency observed in the present study, particularly within high-tempo football contexts demanding rapid decision-making and technical precision. A further applied consideration is that VR-based training is cognitively demanding, raising the question of whether mental fatigue might moderate the transfer of perceptual–technical gains to on-field performance. Recent evidence shows that acute mental fatigue can impair reaction time (~ 46%) and fine-motor strength (~ 5%), while leaving lower-limb power relatively unaffected [31]. This suggests that decision speed and precision-central to scanning and passing-are particularly vulnerable to cognitive overload. At the same time, brain endurance training interventions have been shown to improve passing, shooting, and vigilance specifically under fatigued conditions in professional players [32]. Our RCT provides the first evidence that immersive VR can enhance perceptual-technical skills in youth football under fresh conditions, while the next logical step is to examine whether these improvements persist, or even adapt, under fatigue states that better mirror match realities. For applied settings, practical monitoring strategies such as brief psychomotor vigilance tests, VAS ratings of mental fatigue, or careful workload sequencing could help coaches manage cognitive load and optimize transfer. From a coaching perspective, the observed improvements suggest that VR-based training can be used to reinforce tactical objectives linked to elite-level performance thresholds. Enhanced scanning and one-touch passing skills not only improve technical execution but also enable teams to stabilize possession through purposeful backward distribution, accelerate combination play to raise completed pass counts, and better manage positional risk-reward decisions. Integrating VR sessions into regular training routines may therefore help young players develop the perceptual–cognitive foundations required to achieve the passing efficiency and circulation patterns associated with success in elite competitions. This study has some limitations that should be considered when interpreting the findings. The sample was relatively small and drawn from a single youth academy, which may limit the generalizability of the results. Nevertheless, post-hoc power analysis indicated that the design was sufficiently powered to detect large effects, while smaller effects may require larger samples to reach significance. This is consistent with the observed pattern of results and highlights the robustness of the significant findings reported. In addition, the intervention period was short, and we did not assess whether the observed improvements transfer to actual match performance. Moreover, retention effects and transferability to competitive match play were not assessed, constraining conclusions regarding long-term efficacy. In addition, while the VR tasks were designed to be football-specific, they could not fully capture the physical and tactical demands of real gameplay. Although blinding of participants and coaches was not feasible due to the nature of the intervention, coders were blinded to group allocation and coding was based on predefined operational definitions, with inter-rater reliability exceeding ICC > 0.90. Because all videos were anonymized, randomly ordered, and cropped to exclude contextual cues (e.g., VR headset visibility or session identifiers), it is highly unlikely that the coders could infer group membership. Additionally, potential expectancy effects cannot be fully ruled out, as players in the VR group may have perceived the training as more novel or engaging, which could have enhanced their motivation and effort independently of the specific intervention. Moreover, since the VR group received additional training exposure compared with the control condition, some of the observed improvements might partially reflect the effects of increased total practice time or heightened engagement rather than the VR-specific perceptual–cognitive load alone. Future studies should therefore include active control groups with equivalent exposure time and engagement levels to better isolate the true effects of immersive VR training. Nevertheless, the present investigation addresses a notable gap in the literature by providing the first empirical evidence on how immersive VR training can simultaneously enhance scanning behavior and passing performance in youth football. These findings offer novel insights into the perceptual–motor mechanisms underpinning technical skill acquisition and highlight the potential of VR as a complementary training modality in modern football development programs. Expanding the scope to tactical decision-making, positioning, and defensive actions, and investigating the underlying neurocognitive mechanisms, would further advance the field. In addition, future VR interventions could be designed to include position-specific scenarios, allowing players to repeatedly practice the high-pressure situations most frequently encountered in their roles. Such targeted training may help reduce error rates in critical match contexts and further enhance the ecological validity of VR-based programs. Conclusion The results suggest that VR-based training may facilitate meaningful improvements in visual exploratory actions and passing efficacy, particularly under time-constrained conditions. These improvements are plausibly mediated by perceptual-cognitive recalibrations and neurophysiological engagement of attentional and visuomotor circuits. While the most pronounced effects were observed in scanning activity and short-distance passing under time pressure, directional gains in other passing categories indicate potential for broader transfer with optimized training stimuli. Collectively, these findings highlight immersive VR technology as a promising adjunct to conventional training regimes, with the potential to accelerate the development of anticipatory decision-making and context-specific motor proficiency in youth football. However, given the small sample size, these results should be interpreted as preliminary and exploratory in nature. Further studies with larger cohorts are warranted to confirm and extend the present findings. Acknowledgements This research, supported by TÜBİTAK under Industrial R&D Program Grant No. 3220145, constitutes part of Kaan Erişik’s master’s thesis. It was also supported by the Eskişehir Technical University as part of the 102-Research Support Project titled “The Effects of Virtual Reality–Based Training in Football on Players’ Performance”, under the supervision of Ali Onur Cerrah. Clinical trial number ClinicalTrials.gov identifier: NCT07144371 (retrospectively registered on 27/08/2025). Abbreviations Virtual reality Authors’ contributions All authors contributed to the study conception and design. Material preparation, data collection, and analysis were performed by K.E., A.O.C, and F.S.C. The first draft of the manuscript was written by K.E., A.O.C, and F.S.C, and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript. Funding No specific funding was received for this study. Data availability The data that support the findings of this study are available from the corresponding author upon reasonable request. The data are not publicly available due to privacy and ethical restrictions. Declarations Ethics approval and consent to participate All the procedures implemented in this study were approved by the Ethics Committee of Eskişehir Technical University (Approval date and number: 17/03/2023, E-87914409-050.03.04-2300016243. Since all participants were under 18 years of age, written informed consent was obtained from their parents or legal guardians after providing detailed information about the study. 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 that support the findings of this study are available from the corresponding author upon reasonable request. The data are not publicly available due to privacy and ethical restrictions.