Effects of stroboscopic visual training on reaction time and movement accuracy in collegiate athletes: a systematic review and meta-analysis

Abstract

This systematic review and meta-analysis evaluated the effects of stroboscopic training on sport-specific performance in collegiate athletes and examined the moderating influence of key training parameters. A comprehensive literature search was conducted across multiple databases. The risk of bias in the included studies was assessed using the Cochrane Risk of Bias tool. A random-effects model was applied for the meta-analysis, with separate analyses conducted for time-based and accuracy-based performance measures. Subgroup analyses investigated the moderating effects of training duration, frequency, session length, strobe frequency, duty cycle, athlete age, and training experience. Twelve randomized controlled trials involving 415 athletes met the inclusion criteria. The meta-analysis demonstrated that stroboscopic training had significant moderate effects on both time-based outcomes (SMD = -0.61, p = 0.045) and accuracy-based outcomes (SMD = 0.73, p < 0.01). Subgroup analyses suggested that training protocols lasting 6–10 weeks, with 2–3 sessions per week and 10–20 min per session, were more effective in enhancing overall athletic performance. In terms of strobe parameters, frequencies of 5–20 Hz and duty cycles of 50–70% and 10–50% were associated with greater improvements in accuracy-based and time-based outcomes, respectively. Furthermore, adolescent athletes (10–18 years) exhibited more substantial gains in time-based performance, while those with 4–8 years of training experience achieved the greatest overall improvements. Stroboscopic training effectively enhances sport-specific performance in collegiate athletes. Protocols of 6–10 weeks, incorporating 2–3 weekly sessions of 10–20 min and strobe frequencies of 5–20 Hz, appear to provide the most comprehensive benefits. Optimal duty cycles may differ depending on performance type—50–70% for time-based outcomes and 10–50% for accuracy-based outcomes. These findings offer preliminary guidance for the practical application of stroboscopic training in athletic populations. However, given the limitations in study quality and sample size, further research is warranted to identify the most effective training configurations.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-10393-4.

Introduction

With the continual advancement of competitive sports and the increasing intensity of gameplay, athletes are under growing pressure to process visual information in real time. For instance, a survey of international professional tennis competitions revealed that, over the past decade, the average serve speed in men’s matches has increased from approximately 160 km/h to 220 km/h. In such high-speed environments, even a 0.02-second delay in reaction time can lead to a reduction of up to 10% in return success rate¹. Similarly, in physically demanding sports such as basketball and football, critical performance metrics—such as shooting accuracy—are strongly linked to reaction speed. Athletes who cannot rapidly perceive and anticipate opponents’ actions are more likely to miss defensive opportunities or fail to capitalize on offensive plays². Thus, in today’s high-intensity competitive landscape, insufficient speed and precision in visual perception and response significantly compromise athletic performance³.

In recent years, a novel perceptual training approach—Stroboscopic Visual Training (SVT)—has garnered growing interest within the field of sports science. SVT typically employs specialized eyewear (e.g., Nike Strobe or Senaptec Strobe) that alternates between transparent and opaque lens states at controlled frequencies, thereby creating an environment of intermittent visual occlusion in which participants are required to complete designated motor tasks⁴. Grounded in perception–cognition theory, SVT is hypothesized to enhance the brain’s predictive processing under conditions of partial visual input, while also improving attentional control and visuomotor coordination. These adaptations are believed to translate into improved motor performance under normal visual conditions⁵. Neurophysiological studies have shown that SVT facilitates signal transmission along the dorsal visual stream, primarily via magnocellular pathways. It promotes phase resetting of alpha oscillations in the occipital cortex and supports cross-frequency coupling between alpha and gamma rhythms, while also enhancing predictive coding activity in the parietal cortex⁶. Evidence from EEG studies indicates that 4–6 weeks of SVT can significantly shorten the P100 latency in visual evoked potentials (by 6–12 ms, p = 0.010, ηp² = 0.21) and increase VEP amplitude, suggesting accelerated and intensified early-stage processing of dynamic visual input⁷. Additionally, during stroboscopic balance tasks, participants exhibited reduced visual dependency and enhanced beta-band connectivity within the sensorimotor system. These findings suggest that SVT may improve motor control efficiency through multisensory reweighting mechanisms⁸.

In the field of sports training, stroboscopic visual training (SVT) has gained increasing recognition in recent years and is now progressively being integrated into high-performance athletic development programs. According to a recent review by Jothi, Dhakshinamoorthy and Kothandaraman⁹, at least 74 English-language studies published between 2010 and 2024 have examined the application of SVT in sport-specific contexts. Notably, research activity has surged since 2018, reflecting growing interest in the role of SVT in enhancing athletic performance across various sporting disciplines. The majority of these studies have focused on ball sports. For example, Appelbaum, Cain, Schroeder, Darling and Mitroff¹⁰ were among the first to incorporate Nike stroboscopic eyewear into training protocols for soccer and basketball players. Their findings demonstrated that both stroboscopic training groups (immediate and delayed retest) showed significant improvements in short-term memory retention (p < 0.05); even with extended stimulus–cue intervals (640–2560 ms), participants were able to complete recall tasks effectively. Similarly, Schootemeijer and Visch¹¹ conducted a six-session SVT intervention with tennis athletes and found that the experimental group achieved a rally success rate exceeding 80%, accompanied by a large effect size (Cohen’s d = 0.70). In contrast, the control group exhibited no significant pre–post changes in stroke accuracy (p = 0.34), with a smaller effect size (d = 0.20). These results suggest that short-term SVT may substantially enhance tennis performance, particularly in terms of response precision and perceptual–motor coordination. Extending these findings to adolescent populations, Zwierko, Jedziniak, Popowczak and Rokita¹² reported that a six-week SVT program significantly improved visuomotor response speed—including both reaction time and agility—in youth volleyball players (p = 0.004), yielding a large effect size (d = 0.87). These outcomes further support the potential of SVT to enhance perceptual–motor responsiveness in developing athletes.

Despite the generally positive findings regarding the benefits of stroboscopic visual training (SVT), several studies have reported mixed or limited outcomes, suggesting variability in its effectiveness across different perceptual–cognitive domains. For instance, Ellison, Jones, Sparks, Murphy, Page, Carnegie and Marchant¹³ found that acute SVT—conducted under visually restricted conditions using stroboscopic goggles—significantly improved hand–eye coordination in male volleyball players, with immediate post-training effects (p = 0.003) and sustained improvements observed at both 10 min (p = 0.001) and 10 days (p = 0.002) following the intervention. However, no significant enhancement was observed in visual search ability at the 10-day follow-up (p > 0.05), indicating that SVT may produce domain-specific effects that do not generalize across all perceptual–cognitive functions. Similarly, Fortes, Faro, Faubert, Freitas-Júnior, Lima-Junior and Almeida¹⁴, in a randomized controlled trial, reported that professional soccer players improved their accuracy in predicting pass or shot direction during video-based decision-making tasks (p < 0.05), with the SVT group showing significantly greater gains than the control group (p < 0.05). These findings suggest that SVT may be particularly effective in enhancing anticipatory decision-making. Nevertheless, its capacity to improve higher-order perceptual–cognitive abilities remain inconclusive and warrants further investigation. Additionally, Strainchamps, Ostermann, Mierau and Hülsdünker¹⁵ examined 28 international-level table tennis players and found that following a 10-minute sport-specific warm-up performed while wearing stroboscopic glasses, participants exhibited a significant reduction in reaction time (p < 0.001, η²ₚ = 0.393), yet no significant change was observed in stroke timing (p = 0.246, η²ₚ = 0.055). These results indicate that while SVT may acutely enhance visuomotor response speed, its short-term application does not appear to provide additional benefits over standard visual conditions in the context of fine motor timing.

In summary, Stroboscopic visual training (SVT), a perceptual–cognitive intervention designed to enhance athletic performance, has received increasing empirical and practical attention in competitive sports. Numerous studies suggest that SVT may improve reaction time, hand–eye coordination, and anticipatory skills. However, its overall efficacy remains inconclusive due to substantial variability in participant populations, intervention designs, and device parameters, which has limited the comparability and generalizability of findings across studies. Despite the growing number of empirical investigations into SVT within sport contexts, systematic and quantitative evaluations are still lacking, and the field remains largely unexplored in terms of meta-analytic synthesis. To date, no unified conclusions have been reached regarding the magnitude or scope of SVT’s effects on athletic performance, and its mechanisms of action remain insufficiently defined. This highlights the urgent need for comprehensive, evidence-based evaluation to clarify the overall impact of SVT on various dimensions of sport-specific performance and to explore potential moderating mechanisms. Despite the growing number of empirical investigations into SVT within sport contexts, systematic and quantitative evaluations are still lacking, and the field remains largely unexplored in terms of meta-analytic synthesis. To date, no unified conclusions have been reached regarding the magnitude or scope of SVT’s effects on athletic performance, and its mechanisms of action remain insufficiently defined. This highlights the urgent need for comprehensive, evidence-based evaluation to clarify the overall impact of SVT on various dimensions of sport-specific performance and to explore potential moderating mechanisms. Given that SVT is most frequently applied in ball sports requiring high temporal precision and visuomotor accuracy—and that collegiate athletes are among the primary target populations for such interventions⁹ —this study is both timely and practically relevant. The present research adopts a systematic review and meta-analytic approach to evaluate the effectiveness of SVT in enhancing time-based and accuracy-based sport performance among collegiate athletes. Furthermore, it explores the moderating roles of key intervention characteristics (e.g., training frequency, duration, and cycle) and device-specific parameters (e.g., strobe frequency, duty cycle) in influencing training efficacy. The overarching aim is to provide both theoretical insight and practical guidance for the development of empirically grounded SVT protocols in athletic training environments.

Materials & methods

This study was conducted in strict accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines¹⁶, with all required elements reported accordingly (see Supplementary Checklist: https://www.crd.york.ac.uk/PROSPERO2/view/CRD42025631186). The review protocol was prospectively registered in the PROSPERO database (Registration number: CRD42025631186; January 16, 2025). It should be noted that several items listed in the original PRISMA registration checklist differ from the final methods and analyses reported in this study. These discrepancies reflect methodological adjustments and refinements made during the review process to enhance clarity and ensure alignment with the available evidence. In the interest of transparency and methodological rigor, all deviations from the original protocol—along with their justifications—are fully documented in Appendix 2, available in the Supplementary Materials section of this manuscript.

Literature search

A comprehensive literature search was conducted using the Web of Science, PubMed, and EBSCO databases, covering all records from database inception through December 5, 2024. This systematic review was carried out in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines¹⁶. Domain experts systematically identified relevant keywords and their synonyms with reference to the MeSH database. Keywords included, but were not limited to: “Flicker Training,” “Stroboscopic Training,” “Visual Occlusion Training,” “College Athletes,” “University Athletes,” “Student Athletes,” “Resistance Training,” “Strength Training,” “Muscle Strength,” “Motor Performance,” “Athletic Performance,” “Speed,” “Agility,” “Reaction Time,” “Neuromuscular Control,” “Perceptual-Cognitive Skills,” “Sport-Specific Performance,” “Visual Skill Training,” “Sensorimotor Training,” and “Cognitive Enhancement.” As detailed in Appendix 1 (“Search Strategy”), these terms were combined in various configurations using Boolean operators (AND, OR) and applied uniformly across all databases.

Inclusion and exclusion criteria

The inclusion criteria and core study elements were defined based on the PICOS framework. The specific details are presented in Table 1.

Table 1.

PICOS- inclusion and exclusion Criteria.

Criteria	Inclusion	Exclusion
P (Participants)	Healthy collegiate athletes (professional, semi-professional, or amateur), actively engaged in training or competition	Para-athletes, individuals undergoing injury rehabilitation, or other special populations
I (Intervention)	Use of wearable stroboscopic devices (e.g., Senaptec Strobe, Nike Vapor Strobe, PLATO goggles)	Interventions involving additional forms of sensory stimulation (e.g., auditory cues, tactile feedback, virtual reality, augmented reality, or multimodal perceptual–cognitive training) that may confound the isolated effects of stroboscopic training
C (Comparison)	Control groups using non-stroboscopic settings (e.g., device without strobe function) or conventional visual training	No clearly defined control group, or the control intervention was not comparable to stroboscopic training (e.g., dynamic visual training or multisensory stimulation)
O (Outcomes)	Sport-specific performance outcomes: ·Time-based (movement time, reaction time, anticipation time, reactive agility); · Accuracy-based (success rate, contact accuracy)	Outcomes not directly related to sport-specific performance (e.g., balance, spatial judgment, eye–hand coordination)
S (Study Design)	Randomized Controlled Trials (RCTs), published in English	Non-RCTs, qualitative studies, case reports, reviews, conference abstracts, or incomplete data

Data extraction and coding

The literature search was independently conducted by two reviewers (Luo and Li). Prior to the formal search, both reviewers underwent standardized training to ensure consistency in implementing the search strategy, navigating the databases, and applying the inclusion and exclusion criteria. Upon completion of the initial search, preliminary results were reviewed and cross-verified by Pan and Shi to ensure the accuracy, completeness, and consistency of the process. During the screening phase, titles and abstracts were independently assessed by Luo and Li. For studies requiring full-text review, any discrepancies—such as disagreements regarding study type classification, intervention categorization, or outcome eligibility—were resolved through third-party adjudication by Cao, who also facilitated online discussions. As an expert in systematic review methodology, Cao was responsible for methodological oversight and quality assurance. She made final inclusion decisions based on the original study design, reported outcomes, and alignment with predefined eligibility criteria, thereby ensuring the scientific rigor and consistency of the screening process. The entire literature identification and selection process was conducted in strict accordance with PRISMA guidelines. EndNote 20 was used for reference management to maintain standardization and reproducibility throughout all stages of the review.

A predefined extraction form was employed to systematically record and code all relevant data. The extracted information included: (1) Basic bibliographic details, such as author, country, and year of publication. (2) Participant characteristics, including sample size, age, sex, years of training, competitive experience, and athletic level. (3) Intervention details, encompassing study design, intervention duration, training frequency, type of stroboscopic device, testing instruments, and type of performance outcome (time-based or accuracy-based). (4) Primary outcome measures, specifically time-based and accuracy-based sport-specific performance indicators, reported as mean ± standard deviation. For potential moderator variables—such as intervention duration, frequency, session length, duty cycle, strobe frequency, and participant age—mean values were extracted. (5) Studies reporting multiple time-based or accuracy-based outcomes were treated as containing multiple independent entries. (6) Strobe frequency was classified into low, medium, and high ranges according to predefined criteria. (7) The timing of the stroboscopic intervention was recorded, including whether it was used during warm-up, relaxation, or the primary training phase. For studies presenting outcome data graphically, the Web Plot Digitizer tool (version 4.5; https://apps.automeris.io/wpd/, r = 0.99, p < 0.001) was used to extract values. This method was applied to four studies^12,15,17,18. In cases of missing data, corresponding authors were contacted to obtain the necessary information. If no response was received within 48 h, a follow-up email was sent. Studies were excluded if no reply was received within an additional 48 h.

Risk of bias assessment

The risk of bias in the included randomized controlled trials was systematically assessed using the second version of the Cochrane-endorsed Risk of Bias tool (RoB 2)¹⁹. This tool evaluates potential sources of systematic bias across five predefined domains, encompassing key aspects of trial design, implementation, and reporting: (1)Bias arising from the randomization process; (2)Bias due to deviations from intended interventions; (3) Bias due to missing outcome data; (4)Bias in the measurement of outcomes; and (5) Bias in the selection of the reported results. Each domain includes a structured set of signaling questions to guide reviewers in assigning a domain-level judgment of “low risk of bias,” “some concerns,” or “high risk of bias.” The overall risk of bias for each study is determined by the highest level of risk across all domains—if any single domain is rated as “high risk,” the study is classified as having a high overall risk of bias. Assessments were based on information extracted from full-text articles, supplementary materials, trial registrations, and, where available, additional data provided by the original study authors. Two reviewers (Luo and Li) independently performed the RoB 2 assessments. Discrepancies were resolved through discussion, or, if consensus could not be reached, through consultation with a third reviewer (Cao).

Quality assessment of included studies

The methodological quality of the included randomized controlled trials (RCTs) was assessed using the Physiotherapy Evidence Database (PEDro) scale. The PEDro scale comprises 11 items, 10 of which contribute to the total score. It evaluates key methodological aspects of trial design, including random allocation, allocation concealment, baseline comparability, blinding of participants, therapists, and assessors, completeness of outcome data, intention-to-treat (ITT) analysis, between-group comparisons, and reporting of outcome variability. Each item is rated as either “yes” (1 point) or “no” (0 points), yielding a maximum score of 10. Studies scoring ≥ 6 are considered to be of moderate or higher quality, while those scoring ≥ 8 are classified as high quality. PEDro assessments were independently conducted by two reviewers (Luo and Li). Any scoring discrepancies were resolved through discussion with a third reviewer (Cao), who facilitated consensus to ensure objectivity and consistency in the evaluation process.

Statistical method

All statistical analyses were conducted using Stata version 18.0 and included sensitivity analysis, effect size aggregation, meta-regression, subgroup analysis, and assessment of publication bias. Outcome data were expressed as mean ± standard deviation (Mean ± SD), and 95% confidence intervals (CIs) were calculated for effect sizes and heterogeneity estimates. The Cochrane Q statistic was interpreted based on its p-value rather than confidence intervals, consistent with standard practice. Statistical significance was set at p < 0.05 Heterogeneity among studies was evaluated using the Q-test and the I² statistic. When the Q-test yielded a p-value > 0.1 and I² was less than 50%, heterogeneity was considered low, and a fixed-effects model was applied. In the presence of significant heterogeneity, a random-effects model was employed^20,21. Given the typically small sample sizes in wearable device studies, effect sizes were calculated using Hedges’ g, with interpretation based on the following thresholds: small (SMD < 0.2), moderate (0.2 ≤ SMD < 0.5), large (0.5 ≤ SMD < 0.8), and very large (SMD ≥ 0.8)²². To calculate the standardized mean difference (SMD), M₁ and M₂ denote the pre- and post-intervention means, n₁ and n₂ represent the respective sample sizes, and SD₁ and SD₂ correspond to the standard deviations of the two groups^23,24. The specific formula is presented below. Given the relatively small sample sizes typically observed in SVT studies, this meta-analysis applied a correction factor proposed by Hedges and Tipton²⁵ to adjust for small-sample bias in effect size estimation. When substantial heterogeneity was detected, additional sensitivity analyses, meta-regression, and subgroup analyses were conducted to explore potential sources of heterogeneity. Publication bias was assessed using funnel plots, along with Begg’s and Egger’s statistical tests²⁶. For outcomes involving fewer than 10 studies, the trim-and-fill method was employed to adjust for potential publication bias and enhance the robustness of the findings²⁷. Additionally, to account for the presence of multiple dependent effect sizes within individual studies, the analysis applied a Robust Variance Estimation (RVE) approach with a pre-specified correlation coefficient (ρ) of 0.80, alongside a three-level nested random-effects model with CR2 adjustment²⁸. This combined strategy was implemented to appropriately model within-study dependency and enhance the robustness of the statistical inferences²⁹.

1

2

Results

Literature search results

As illustrated in Fig. 1, A total of 456 studies were initially identified through database searches, including 448 from electronic databases and 8 from additional sources. After removing duplicates using EndNote 19, 35 studies remained for further evaluation. Following title and abstract screening, 398 studies were excluded for not meeting the inclusion criteria. Subsequently, the full texts of 23 studies were assessed in detail, leading to the exclusion of 11 studies due to inappropriate study designs, irrelevant outcome measures, or incomplete data. Ultimately, 12 studies met all inclusion criteria and were included in the final meta-analysis.

Fig. 1.

Flowchart of literature screening.

Characteristics of included studie

As summarized in Table 2, this review included 12 randomized controlled trials (RCTs) examining the effects of stroboscopic visual training on athletic performance. The studies encompassed a range of ball sports—including volleyball, table tennis, soccer, cricket, badminton, and tennis—reflecting a broad representation across athletic disciplines. Participants were primarily adolescent and young collegiate athletes, aged 11 to 32 years, with average ages ranging from 16 to 22 years. Most studies involved athletes with 2 to 7 years of training experience. While male participants predominated, several studies also included female athletes. In the intervention protocols, athletes in the experimental groups engaged in sport-specific training while wearing stroboscopic eyewear. Strobe frequencies typically ranged from 4 to 20 Hz, with duty cycles between 30% and 70%. Training durations spanned 4 to 10 weeks, with sessions held 2 to 3 times per week, each lasting 20 to 30 min. Control groups received equivalent conventional training without stroboscopic visual input. Primary outcome measures included execution time, reaction time, reactive agility, anticipation accuracy, and the success rate of technical skill execution. All included studies adopted randomized controlled designs to evaluate the effects of stroboscopic visual training on athletes’ perceptual–motor performance.

Table 2.

Basic characteristics of included Studies.

Study	Participants	Intervention		Outcomes
		Methods	Duration
Zwierko et al. (2024a)30	50 adolescent volleyball players (24 F, 26 M; age 16.5 ± 0.6 years; training experience: 6.6 years); Exp/Con = 25; RCT	Experimental Group: Wore stroboscopic eyewear during volleyball-specific functional training. Frequency: 15 ~ 9 Hz; Duty cycle: 50%~70%. Each cycle: 5 min (2.5 min rest + 2.5 min training). Control Group: Conventional volleyball training without stroboscopic stimulation.	6 weeks、3 sessions/week、 25 ~ 30 min per session	Successful block execution time (s)
Zwierko et al. (2024b)12	50 adolescent volleyball players (24 F, 26 M; age 16.5 ± 0.6 years; training experience: 6.65 ± 1.2 years); Exp/Con = 25; RCT	Experimental Group: Wore stroboscopic eyewear during volleyball-specific functional training (frequency: 15 ~ 9 Hz; duty cycle: 50%~70%). Each cycle: 5 min (2.5 min rest + 2.5 min training). Control Group: Conventional volleyball training without stroboscopic stimulation.	6 weeks、3 times per week、 25 ~ 30 min per session	Reaction agility (s)
Strainchamps et al. (2023)15	28 international table tennis players (17 F, 11 M; age 21.04 ± 6.4 years; training experience: 12.89 ± 6.8 years); Exp/Con = 14; RCT	Experimental Group: Wore stroboscopic eyewear during table tennis skill training (frequency: 20 Hz; duty cycle: 50%). Each cycle: 3 ~ 4 min (1 min rest + 2–3 min training). Control Group: Conventional table tennis training without stroboscopic stimulation.	2 weeks、2 times per week、 30 min per session	Successful stroke time (s), Stroke reaction time (s)
Zwierko et al. (2023)31	50 adolescent volleyball players (24 F, 26 M; age 16.5 ± 0.6 years; training experience: 6.65 ± 1.2 years); Exp/Con = 25; RCT	Experimental Group: Wore stroboscopic eyewear during volleyball-specific functional training (frequency: 15 ~ 9 Hz; duty cycle: 50%~70%). Each cycle: 5 min (2.5 min rest + 2.5 min training). Control Group: Conventional volleyball training without stroboscopic stimulation.	6 weeks、3 times per week、 25 ~ 30 min per session	Successful smash time (s)、Stroke reaction time (s)、Reaction agility (s)
Fortes et al. (2023)14	28 male professional soccer players (age: 18 ~ 32 years; training experience: 7.5 years); Exp/Con = 14; RCT	Experimental Group: Wore stroboscopic eyewear during small-field soccer training (frequency: 4.76 Hz; duty cycle: 47.62%). Control Group: Wore fixed-frequency stroboscopic eyewear without active flashing.	10 weeks、3 times per week、 24 min per session	Successful anticipation rate (%)、Anticipation reaction time (s)
Palmer et al. (2022)32	36 club-level male cricket batters (age: 11.2 ± 1.3 years); Exp/Con = 18; RCT	Experimental Group: Wore stroboscopic eyewear during dribbling practice (frequency: 1.33 ~ 4 Hz; duty cycle: 13.33%~40%). Control Group: Trained under normal visual conditions.	4 weeks、2 sessions/week、 20 min per session	Successful ball touch time (s), Successful ball touch rate (%)
Hülsdünker et al. (2021a)33	32 elite badminton players (21 F, 11 M; age 13.7 ± 1.3 years; training experience: 6.2 ± 2.2 years); Exp/Con = 16; RCT	Experimental Group: Wore stroboscopic eyewear during individual badminton/shuttlecock machine training (frequency: 15 ~ 8 Hz; duty cycle: 50%~70%). Control Group: Identical badminton training without stroboscopic stimulation.	10 weeks、3 ~ 6 times per week、 10 ~ 15 min/session	Stroke reaction time (s)、Successful stroke time (s)
Hülsdünker et al. (2021b)34	32 elite badminton players (21 F, 11 M; age 13.7 ± 1.3 years; training experience: 6.2 ± 2.2 years); Exp/Con = 16; RCT	Experimental Group: Wore stroboscopic eyewear during badminton/shuttlecock machine training (frequency: 15 ~ 8 Hz; duty cycle: 50%~70%). Control Group: Identical training without stroboscopic stimulation.	10 weeks、3 ~ 6 times per week、 10 ~ 15 min/session	Stroke reaction time (s)、 Successful stroke time (s)
Boiangin (2019)35	30 intermediate tennis players (15 F, 15 M; age 18 ~ 22 years; training experience: 8.27 years); Exp1/Exp2/Con = 10; RCT	Exp1 (Mid-frequency): Wore stroboscopic eyewear during tennis serve training (frequency: 5 Hz; duty cycle: 50%). Exp2 (Low-frequency): Non-stroboscopic eyewear (frequency: 0.56 Hz; duty cycle: 50%). Control Group: Non-stroboscopic eyewear.	1 week、6 times per week、30 min/session	Anticipation accuracy (%)、Anticipation reaction time (s)
Hülsdünker et al. (2019)36	10 club-level male badminton players (1 F, 9 M; age 13.4 ± 3 years); Exp/Con = 5; RCT	Experimental Group: Wore stroboscopic eyewear during badminton training (frequency: 6 ~ 5 Hz; duty cycle: 50%~70%). Control Group: Identical training without stroboscopic stimulation.	4 weeks、3 times per week、25 ~ 30 min/session	Smash success rate (%)
Schootemeijer and Visch (2017)11	19 high-performance tennis players (age 13.66 ~ 13.74 years; training experience: 7.95 ~ 8.12 years); Exp = 10, Con = 9; RCT	Wore stroboscopic eyewear during cricket-specific training (frequency: 7 ~ 5 Hz; duty cycle: 30%~50%). Control Group: Stroboscopic eyewear without flicker (100% clear state).	3 weeks、2 sessions/week、60 ~ 70 min/session	Stroke success rate (%)
Wilkins and Gray (2015)17	30 tennis players (12 F, 18 M; age 22.5 ± 4.7 years; training experience: 9.1 ± 0.1 years); Exp/Con = 15; RCT	Experimental Group: Wore stroboscopic eyewear during tennis training (Prince Professional II at 22 mph; frequency: 0.1 ~ 0.4 Hz; duty cycle: 10%~40%). Control Group: Fixed-frequency stroboscopic eyewear.	6 weeks、2 times per week、20 min/session	Catch success rate (%)

Note: F: Female | M: Male | Exp: Experimental Group | Con: Control Group | mph: Miles per Hour.

Risk of bias assessment

As shown in Fig. 2, the risk of bias in the included studies was assessed using the RoB 2 tool, which comprises five structured domains, and visualized using Review Manager 5.4. The analysis revealed that in approximately half of the studies, the predominant source of bias stemmed from deviations from intended interventions, as observed in the works of Zwierko et al., Palmer et al., and Hülsdünker et al. A smaller subset of studies exhibited concerns regarding potential bias in outcome measurement, including those by Wilkins and Gray, Hülsdünker et al., and Zwierko et al. Overall, nearly all included studies demonstrated at least one or two domains with either high or unclear risk of bias, potentially attributable to the inherent variability and uncontrollability of factors in exercise-based interventions. These findings highlight the need for cautious interpretation of the results and underscore the importance of stricter adherence to experimental protocols and standardized measurement procedures in future research.

Fig. 2.

Risk of bias assessment.

Quality assessment of included studies

As summarized in Table 3, the methodological quality of the included studies was evaluated using the Physiotherapy Evidence Database (PEDro) scale. The results showed that while most studies clearly reported randomization procedures, there was inconsistent or insufficient reporting regarding allocation concealment, blinding methods, and the use of intention-to-treat analysis. Nevertheless, the majority of studies performed satisfactorily in terms of data completeness, appropriate statistical comparisons, and reporting of variability measures. Overall, the included trials demonstrated moderate to high methodological quality, indicating a reasonable degree of design rigor and scientific reliability. However, the limited detail in reporting randomization and blinding procedures warrants caution in interpreting the findings and drawing definitive conclusions.

Table 3.

Methodological quality assessment of included studies (PEDro Scale).

Author and year	2	3	4	5	6	7	8	9	10	11	总分
Zwierko et al. (2024a)	✓	✗	✓	?	?	?	✓	✓	✓	✓	6
Zwierko et al. (2024b)	✓	✗	✓	?	?	?	✓	✓	✓	✓	6
Strainchamps et al. (2023)	✗	?	✓	✗	✗	✗	✓	✓	✓	✓	5
Zwierko et al. (2023)	✓	✗	✓	?	?	✓	✓	✓	?	✓	6
Fortes et al. (2023)	✓	✗	✓	✗	?	✗	✓	✓	✓	✓	6
Palmer et al. (2022)	✗	✗	✓	✗	?	✗	✓	✓	✓	✓	5
Hülsdünker et al. (2021a)	✓	?	✓	?	✗	✗	✓	✗	✓	✓	5
Hülsdünker et al. (2021b)	✓	?	✓	?	✗	✗	✓	✗	✓	✓	5
Boiangin (2019)	✓	✗	✓	✗	✗	?	✓	✓	✓	✓	6
Hülsdünker et al. (2019)	✓	?	✓	?	✗	✗	✓	✗	✓	✓	5
Schootemeijer et al. (2017)	?	✗	✓	✗	✗	✗	✓	✓	✓	✓	5
Wilkins and Gray (2015)	✓	✗	✓	?	✗	✓	✓	✗	✓	✓	6

Note: Items 2 through 11 of the PEDro scale were used for scoring, with each “Yes” response receiving 1 point and a maximum total score of 10. 2. Random allocation;3. Concealed allocation. 4. Baseline comparability. 5. Participant blinding. 6. Therapist blinding. 7. Assessor blinding. 8. Adequate follow-up. 9. Intention-to-treat analysis. 10. Between-group comparisons. 11. Reporting of point estimates and variability. A score of ≥ 6 was considered moderate-to-high quality, and ≥ 8 was considered high quality. A checkmark “✓” indicates “Yes,” a cross “✗” indicates “No,” and a blank or “?” indicates “Unclear.”.

Meta-analysis

Heterogeneity and sensitivity analysis

According to the heterogeneity test results presented in Table 4, both time-related and accuracy-related sports performance indicators exhibited moderate heterogeneity among the included athlete studies (I²>50%, p < 0.1). Considering the characteristics of these studies, it is hypothesized that variations in outcome indicators may be the primary source of heterogeneity. To validate this hypothesis, a sensitivity analysis was conducted to further confirm the robustness of the findings.

Table 4.

Heterogeneity analysis on exercise effects of included studies.

Outcome Category	Q	I2 (%)	P	SMD	CI	Z	P
Time-related (s)	39.78	61.83	0.00	−0.59	−0.87, −0.31	−4.15	0.00
Accuracy-related (%)	27.09	68.00	0.00	0.73	0.17, 1.28	2.57	0.01

As illustrated in Fig. 3, the included studies were categorized into two groups: one focusing on time-related sports performance indicators and the other on accuracy-related sports performance indicators. A leave-one-out sensitivity analysis was conducted to assess the influence of individual studies on the overall effect size. This method involved sequentially excluding each included study and recalculating the standardized mean difference (SMD) and its 95% confidence interval (CI) using a random-effects model³⁷. A study was considered to have a substantial impact if its exclusion resulted in one or more of the following: (1) a reversal in the direction of the overall effect; (2) a 95% CI that crossed the null value; (3) a change in the I² statistic exceeding 10%; or (4) a shift in statistical significance (i.e., from significant to non-significant, or vice versa). In such cases, the study was excluded from the final synthesis, and the pooled effect size was recalculated accordingly²¹. The findings revealed that the point estimates of the overall effect size fluctuated within ± 5%, with no significant variations. The width of the confidence intervals (CIs) remained within ± 5%, with consistent upper and lower bound trends that did not cross zero. Additionally, the heterogeneity index (I²) varied by less than ± 10%, indicating no substantial change in heterogeneity classification. The statistical significance (p-value) also remained unchanged following the exclusion of any single study, confirming the stability of the results. In conclusion, the sensitivity analysis demonstrated robust findings, suggesting that the impact of outcome indicators on heterogeneity is minimal. Instead, the observed heterogeneity across studies is likely attributable to other factors beyond the classification of outcome indicators.

Fig. 3.

Sensitivity analysis on exercise effects of included studies.

Pooled effect size

This study included 13 studies, of which 10 reported time-related sports performance indicators and 7 reported accuracy-related sports performance indicators. Given the distinct performance characteristics of these two types of outcome indicators, separate analyses were conducted to ensure a precise evaluation of their effects.

As illustrated in Fig. 4, this meta-analysis included eight independent studies, several of which reported multiple dependent outcomes—specifically, Zwierko et al. (5 outcomes), Hülsdünker et al. (5), Strainchamps et al. (2), and Boiangin (2019) (2). To account for the dependency among effect sizes within studies, robust variance estimation (RVE) with a weighted regression model was employed, assuming an intra-class correlation coefficient (ρ) of 0.80. The main analysis revealed that stroboscopic visual training (SVT) exerted a moderate negative effect on time-based performance outcomes, with an effect size of g = − 0.61, SE = 0.24, 95% CI [–1.24, 0.01], t (df = 5) = − 2.54, p = 0.045. Sensitivity analyses using alternative ρ values (0.60 and 1.00) yielded consistent results (g ranging from − 0.59 to − 0.63), with minimal fluctuation (< 0.03) and no meaningful change in statistical significance. These findings confirm the robustness of the results to assumptions regarding within-study correlations (see Appendix 3). In addition, a three-level random-effects model with CR2 adjustment produced a comparable effect size (g = − 0.58, 95% CI [–1.23, 0.07]). Variance decomposition indicated that approximately 77% of the total heterogeneity was attributable to between-study differences (τ²_between = 0.20), with the remaining variance arising from within-study heterogeneity (τ²_within = 0.06). Taken together, the results from the primary RVE analysis, sensitivity tests across correlation assumptions, and three-level modeling suggest that SVT may yield a moderate beneficial effect—equivalent to approximately 0.6 standard deviations—in time-sensitive athletic performance. Nevertheless, the relatively small sample size and a confidence interval that includes the null warrant cautious interpretation and underscore the need for further validation through rigorously designed randomized controlled trials.

Fig. 4.

Three-level random-effects forest plot of the effects of stroboscopic training on time-dependent performance outcomes (in seconds) among adolescent athletes.

As illustrated in Fig. 5, the heterogeneity of accuracy-related sports performance indicators was relatively high (I² = 68.00%, p < 0.01). The pooled effect size (SMD = 0.73, p < 0.01) indicated a moderate effect, which was statistically significant.

Fig. 5.

Forest Plot of the Meta-Analysis on the Effects of Stroboscopic Training on Accuracy-Related (%) Performance Variables in Adolescent Athletes.

Subgroup analysis

Subgroup analysis of time-related (s) performance indicators

As illustrated in Table 5, The subgroup analysis results indicate that for time-related performance indicators, a training period of 6–10 weeks produced the most significant improvement (SMD = −1.18, p < 0.05). Training for 1–6 weeks resulted in some improvement but did not reach statistical significance (SMD = −0.74, p > 0.05). Training 3–6 times per week significantly enhanced performance (SMD = −0.66, p < 0.05), whereas training less than 3 times per week was not statistically significant (SMD = −0.25, p > 0.05). The optimal training duration per session was 10–20 min, yielding the most significant effect (SMD = −0.86, p < 0.01), while extending sessions to 20–30 min resulted in a weaker but still significant improvement (SMD = −0.40, p < 0.01). Training within the 5–20 Hz range had the most significant effect (SMD = −0.63, p < 0.01), whereas training in the 1.33–5 Hz range showed some improvement but was not statistically significant (SMD = −0.63, p > 0.05). A 50-70% duty cycle yielded the best performance gains (SMD = −0.73, p < 0.01) while reducing the duty cycle to 13-50% led to limited and non-significant improvements (SMD = −0.45, p > 0.05). Adolescents aged 10–18 years showed the most significant improvements (SMD = −0.66, p < 0.05), whereas adult athletes (18–32 years old) did not experience statistically significant benefits (SMD = −0.54, p > 0.05). Athletes with 4–8 years of training experience exhibited the most improvement (SMD = −0.78, p < 0.05), while those with 8–19 years of training experience also showed significant gains (SMD = −0.62, p < 0.01).

Table 5.

Subgroup analysis of Time-Related (s) performance indicators.

Outcomes			Studies (Sample Size)	Meta-Analysis results
				SMD	(95%CI)	Z	P
Time-related (s)	Intervention Duration	1 ~ 6 weeks	3(92)	−0.74	−1.08, 0.39	−4.21	0.00
		6 ~ 10 weeks	6(230)	−1.18	−1.57, −0.79	−5.88	0.00
	Intervention Frequency	2 ~ 3 sessions	3(92)	−0.25	−0.73, 0.23	−1.01	0.23
		3 ~ 6 sessions	7(230)	−0.66	−1.34, −0.41	−3.67	0.00
	Session Duration	10 ~ 20 min	4(146)	−0.86	−1.37, −0.35	−3.28	0.00
		20 ~ 30 min	5(168)	−0.40	−0.68, −0.12	−2.79	0.01
	Stroboscopic Frequency	1.33 ~ 5HZ	3(84)	−0.63	−1.69, 0.44	−1.16	0.25
		5 ~ 20HZ	6(230)	−0.63	−0.95, −0.32	−3.39	0.00
	Duty Cycle	50 ~ 70%	4(182)	−0.73	−1.10, −0.36	−3.88	0.00
		13.33 ~ 50%	4(112)	−0.45	[0.99, 0.10	−1.62	0.11
	Participant Age	10 ~ 18 years	6(250)	−0.66	−1.02, −0.31	−3.62	0.00
		18 ~ 32 years	3(76)	−0.54	−1.18, 0.09	−1.67	0.09
	Training Experience	4 ~ 8 years	3(128)	−0.78	−1.44, −0.13	−2.35	0.02
		8 ~ 19 years	5(162)	−0.62	−0.98, −0.25	−3.34	0.00

Subgroup analysis of accuracy-related performance indicators

As illustrated in Table 6, The subgroup analysis results for accuracy-related performance indicators revealed significant variations in effectiveness under different intervention conditions. Training duration of 6–10 weeks produced the most significant improvement (SMD = 1.02, p < 0.05), whereas training for 1–6 weeks resulted in some improvement but did not reach statistical significance (SMD = 0.55, p > 0.05). Training 3–6 times per week significantly enhanced performance (SMD = 0.74, p < 0.05), while training fewer than 3 times per week was not statistically significant (SMD = 0.72, p > 0.05). The optimal training duration per session was 10–20 min, yielding the most significant effect (SMD = 0.58, p < 0.01), whereas extending sessions to 20–70 min led to a weaker and non-significant effect (SMD = 0.77, p > 0.05). Training within the 5–20 Hz range had the most significant effect (SMD = 1.07, p < 0.05), whereas training in the 1.33–5 Hz range showed diminished effects that were not statistically significant (SMD = 0.53, p > 0.05). A 10-50% duty cycle yielded the most improvement (SMD = 0.93, p < 0.05), whereas increasing the duty cycle to 50-70% resulted in limited and non-significant improvements (SMD = 0.44, p > 0.05). Neither adolescents (11–18 years) (SMD = 0.79, p > 0.05) nor adults (18–32 years) (SMD = 0.47, p > 0.05) demonstrated statistically significant improvements. Athletes with 8–17 years of training experience exhibited the most significant improvement (SMD = −0.78, p < 0.05), whereas those with 7.5-8 years of training experience showed limited and non-significant improvements (SMD = −0.62, p > 0.05).

Table 6.

Subgroup analysis of accuracy-related performance indicators.

Outcomes			Studies (Sample Size)	Meta-Analysis results
				SMD	(95%CI)	Z	P
accuracy-related (%)	Intervention Duration	1 ~ 6 weeks	4(86)	0.55	−0.21, 1.30	1.42	0.15
		6 ~ 10 weeks	3(78)	1.02	0.13, 1.90	2.57	0.01
	Intervention Frequency	3 ~ 6 sessions	4(98)	0.72	−0.52, 1.95	1.14	0.25
		2 ~ 3 sessions	3(86)	0.74	0.10, 1.37	2.27	0.02
	Session Duration	10 ~ 20 min	2(66)	0.58	−1.50, 2.66	0.55	0.00
		20 ~ 70 min	5(118)	0.77	0.25, 1.30	2.90	0.58
	Stroboscopic Frequency	0.1 ~ 5HZ	4(114)	0.53	−0.33, 1.38	1.20	0.23
		5 ~ 20HZ	2(30)	1.07	0.31, 1.8	2.75	0.01
	Duty Cycle	10 ~ 50%	3(78)	0.93	0.08, 1.78	2.15	0.03
		50 ~ 70%	3(66)	0.44	−0.50, 1.38	0.92	0.36
	Participant Age	11 ~ 18 years	3(98)	0.79	−0.11, 1.69	1.72	0.04
		18 ~ 32 years	3(66)	0.47	−0.54, 1.47	0.91	0.36
	Training Experience	7.5 ~ 8 years	2(48)	1.04	0.25, 1.83	2.59	0.01
		8 ~ 17 years	3(60)	0.55	−0.30, 1.40	1.27	0.20

Publication bias

As illustrated in Fig. 6, the funnel plots for time-based performance outcomes (Fig. 6A) and accuracy-based outcomes (Fig. 6B) display a reasonably symmetrical distribution of effect sizes around the central vertical axis, with no prominent outliers observed. This visual pattern suggests a low risk of publication bias in the current meta-analysis and supports the stability and reliability of the pooled effect estimates. Furthermore, for sport-specific accuracy outcomes, Egger’s test yielded Z = 1.53, p = 0.126, and Begg’s test yielded Z = 1.05, p = 0.293. For sport-specific time-based outcomes, Egger’s test yielded Z = − 1.26, p = 0.206, while Begg’s test yielded Z = − 1.54, p = 0.124. These findings indicate no statistically significant evidence of publication bias in either outcome category. confirming the robustness of the study results.

Fig. 6.

Funnel plot for publication bias in included studies.

Discussion

This study provides robust evidence that stroboscopic visual training (SVT) significantly enhances sport-specific performance. Specifically, SVT was associated with moderate-to-large improvements in both time-based performance outcomes (SMD = − 0.61, p = 0.045) and accuracy-based outcomes (SMD = 0.73, p = 0.01).Intervention duration emerged as a key moderator: training programs lasting 6 to 10 weeks yielded the greatest benefits, with effect sizes of SMD = − 1.18 (p = 0.00) for time-based tasks and SMD = 1.02 (p = 0.01) for accuracy-based tasks.Weekly training frequency also played an important role. A frequency of 3 to 6 sessions per week was most effective for enhancing time-based performance (SMD = − 0.66, p = 0.00), whereas 2 to 3 sessions per week resulted in optimal gains in accuracy-based outcomes (SMD = 0.74, p = 0.02).With regard to single-session duration, interventions lasting 10 to 20 min per session were associated with significant improvements in both time-based (SMD = − 1.37, p = 0.00) and accuracy-based outcomes (SMD = 0.58, p = 0.00). In terms of strobe device parameters, flicker frequencies between 5 and 20 Hz were most effective, producing effect sizes of SMD = − 0.63 (p = 0.00) for time-based and SMD = 1.07 (p = 0.01) for accuracy-based performance. A duty cycle of 50–70% yielded the greatest improvements in time-based performance (SMD = − 0.63, p = 0.00), while a lower duty cycle of 10–50% was more beneficial for accuracy-based tasks (SMD = 0.93, p = 0.03). Notably, adolescent athletes aged 11–18 years demonstrated the most pronounced improvements (SMD = 0.66, p = 0.00 for time-based; SMD = 0.79, p = 0.04 for accuracy-based outcomes). Additionally, athletes with 7–8 years of training experience showed substantial gains in both domains (SMD = 0.78, p = 0.02 for time-based; SMD = 1.04, p = 0.01 for accuracy-based performance).

The preliminary meta-analytic findings of this study align broadly with those of several prior investigations. For example, Wilkins and Appelbaum³⁸, in a review of 12 recent publications on stroboscopic visual training (SVT) in sports contexts, concluded that SVT is a promising perceptual–response training tool with the potential to enhance sport-specific performance and transfer training gains to competitive scenarios under normal visual conditions. Similarly, Jothi, Dhakshinamoorthy and Kothandaraman⁹, in a review of 13 studies, reported that SVT significantly improved visual and visuomotor functions, reaction time, and anticipatory hand–eye coordination. However, they also noted adverse effects in some cases, including mental fatigue and difficulty in color recognition. Additional insights have been provided by several individual studies. For instance, Wilkins, Nelson and Tweddle³⁹ implemented a 7-week SVT program with six elite youth soccer goalkeepers, reporting improvements in visual response time but no notable benefits in broader perceptual and cognitive functions, such as attention, anticipatory judgment, or hand–eye coordination. Bennett, Ashford, Rioja and Elliott⁴⁰ observed significant pre–post improvements in participants who practiced single-handed catching under stroboscopic conditions. In another study, Smith and Mitroff⁴¹ investigated the effect of SVT on predictive timing—the ability to estimate the position of a moving object over time—and found that the stroboscopic group responded more quickly and accurately than the control group both immediately post-training and 10 min later.

The observed improvements in sport-specific reaction time and accuracy performance (SMD ≈ 0.6–0.7) are likely underpinned by neurophysiological mechanisms activated through stroboscopic visual training (SVT). Previous research has demonstrated that SVT enhances the efficiency of the dorsal visual stream, particularly through magnocellular-dominated pathways and occipito-parietal cortical projections^7,42. The intermittent visual occlusion characteristic of SVT forces individuals to rely more heavily on predictive coding and short-term visual memory to maintain target tracking and anticipate motion trajectories^43,44. Electrophysiological studies further reveal that such perceptual disruptions lead to a shortening of the P100 latency in visual evoked potentials (VEP-P100) and an increase in alpha–gamma cross-frequency coupling within the occipito-parietal network. These changes suggest faster sensory gating and enhanced top-down modulation of action-relevant stimuli⁴⁵. Simultaneously, increased beta-band connectivity within sensorimotor circuits has been associated with multisensory reweighting processes that may refine visuomotor timing and promote more stable, consistent motor output under dynamic or unpredictable conditions^46,47.

Nevertheless, the effects of SVT are not uniformly significant across all contexts. For example, Strainchamps, Ostermann, Mierau and Hülsdünker¹⁵ reported no significant improvements in reaction speed among elite table tennis players when SVT was applied solely during warm-up. Likewise, Boiangin³⁵ examined the effects of low- and medium-frequency SVT on tennis serve anticipation following a single 30-minute session. Performance was assessed across four time points—pre-intervention, during, immediately post-intervention, and one week later—measuring anticipatory accuracy, reaction time, and self-efficacy. No significant group differences were observed at any time point, suggesting limited impact under those conditions. The limited behavioral improvements observed in some SVT interventions may be attributable to insufficient overall stimulation dosage—specifically, intervention durations, frequencies, or session lengths that fall below the threshold required to trigger neuroplastic adaptation. Neurophysiological evidence suggests that long-term potentiation (LTP) and synaptogenesis—both essential for perceptual–motor learning—depend on repeated, high-intensity stimulation to initiate intracellular signaling cascades, including calcium influx, CaMKII autophosphorylation, and activation of the BDNF–TrkB pathway⁴⁸.Animal studies have shown that approximately 180 min of cumulative stimulation, delivered intermittently, are necessary to induce measurable structural remodeling in regions such as the hippocampus and motor cortex. These changes include AMPA receptor insertion and dendritic spine proliferation⁴⁹. Complementary findings from human studies using transcranial magnetic stimulation (TMS) and neuroimaging indicate that interventions spanning at least 4 to 6 weeks, with a minimum frequency of three sessions per week, are required to produce stable increases in motor-evoked potential (MEP) amplitudes and to enhance α–γ cross-frequency neural coupling⁵⁰. Accordingly, when SVT protocols involve only a single session lasting less than 30 min—or as brief as 10 min—the resulting cortical desynchronization and transient release from inhibition may be limited to short-lived phase-resetting effects. These effects are likely insufficient to surpass the biological threshold necessary for initiating gene transcription or long-term structural remodeling. This may help explain the absence of significant improvements in behavioral outcomes such as reaction time or task accuracy in certain SVT studies³³.These findings highlight the critical role of training protocol design in determining the effectiveness of SVT. When stroboscopic stimuli are applied too briefly or not sufficiently integrated into sport-specific skill training, the potential for meaningful adaptation may be diminished. This interpretation is supported by the subgroup analysis in the present study, which found that SVT had limited effects on accuracy when intervention duration was short (≤ 2 weeks) or when individual session length was insufficient (≤ 10 min). Furthermore, excessively low strobe frequencies or duty cycles may render tasks overly difficult, thereby compromising training quality.

Although the present study did not directly examine the neural or cognitive mechanisms underlying the effects of SVT, previous research has provided converging evidence that SVT may modulate brain activity patterns, enhance perceptual–motor integration, and strengthen attentional control networks. The mechanisms underlying stroboscopic training are primarily associated with adaptations in visual information processing and motor control. SVT imposes a “visual bottleneck” by restricting continuous visual input, compelling athletes to perform movements under conditions of intermittent visual occlusion. This constraint encourages more efficient utilization of limited visual cues and fosters reliance on anticipatory strategies to complete motor tasks⁵¹. When normal visual conditions are restored, these adaptations may result in an “overcompensation” effect, manifesting as enhanced performance. Carrolla, Fullerb, Lawrenceb, Osborne, Stallcup, Burch, Freemand, Chandere, Strawderman and Cranef⁵² emphasized that SVT improves sensitivity to rapidly appearing and disappearing stimuli and enhances the short-term retention of transient visual information. These perceptual–cognitive improvements may translate effectively to competitive sport contexts by enabling athletes to better detect critical cues and make faster, more accurate decisions⁵³. Additionally, SVT requires athletes to maintain attention and motor coordination under heightened perceptual demands, potentially strengthening visuomotor coupling⁷. Upon returning to normal visual conditions, athletes may perceive the environment as clearer and more information-rich, thereby facilitating smoother and more effective performance. This interpretation is supported by climbing-related research^42,54. For instance, Vasile and Stănescu⁴ found that SVT led to greater improvements in climbers’ spatial awareness than in basic reaction time, suggesting that SVT primarily enhances the cognitive components of visual–motor coordination.

When designing SVT protocols, it is essential to recognize that training duration and frequency may differentially influence various aspects of athletic performance. For speed-related outcomes such as reaction time, the present study found that even short-term, low-frequency SVT can produce meaningful improvements. However, extending the training duration does not appear to yield linear benefits, and overly frequent sessions may reduce effectiveness due to accumulated fatigue. These findings suggest that coaches can optimize reaction speed through brief, high-quality sessions conducted 2–3 times per week, each lasting 15–20 min. In contrast, improvements in accuracy-related metrics and technical consistency tend to emerge under conditions of longer-duration and higher-frequency training. This implies that when the objective is to enhance skills such as hitting accuracy or shooting precision, SVT may need to be implemented more frequently and over an extended period to allow for adequate motor adaptation and refinement. In applied practice, SVT protocols can be tailored to align with specific training goals. For instance, during the preparatory phase of a season, high-frequency SVT may be used over several weeks to improve ball-handling precision. Conversely, during the pre-competition phase, a moderate training frequency may be more appropriate to maintain and consolidate reaction responsiveness without inducing fatigue.

In stroboscopic training, excessively low strobe frequencies or duty cycles result in the lenses remaining opaque for the majority of the time, effectively placing athletes in near-total visual darkness. This can distort movement patterns or hinder the proper execution of technical skills, thereby reducing training efficacy^55,56. Conversely, excessively high strobe frequencies or duty cycles closely mimic normal visual conditions and may fail to introduce sufficient visual perturbation to elicit meaningful perceptual–motor adaptation. Therefore, a moderate level of visual challenge is essential—athletes should experience noticeable visual disruption without being entirely deprived of critical visual input. Based on the findings of the present study, an optimal training range appears to be a strobe frequency of approximately 5–15 Hz and a duty cycle around 50%. These parameters have direct implications for training design. When implementing stroboscopic eyewear, coaches are advised to begin with moderate occlusion settings and adjust parameters progressively based on individual adaptation. If the task is overly difficult, the challenge can be reduced by increasing the duty cycle (i.e., extending the transparent phase) or raising the frequency. Conversely, if the visual demands are too low, reducing the duty cycle or lowering the frequency can intensify the stimulus and enhance the training effect.

With regard to individual differences, this study found that adolescent athletes experienced greater benefits from stroboscopic training compared to their adult counterparts. This may be attributed to the fact that adolescents are in a critical phase of neural development and skill acquisition, during which visual training is more likely to produce measurable improvements in performance^2,57. Hülsdünker, Gunasekara and Mierau³³. Supporting this notion, reported neuroplastic changes in brain function six weeks after training in young athletes, suggesting that SVT may induce more profound and lasting adaptations during adolescence. In contrast, adult athletes typically demonstrate more stable sensorimotor systems as a result of long-term training, with reduced neuroplasticity and consequently a more limited capacity for further improvement^58,59. Similarly, athletes with less training experience tended to exhibit greater gains following SVT than those with more extensive experience, consistent with the well-established “novice gain” phenomenon^60,61. For novice athletes, SVT may accelerate the development of perceptual–decision-making skills. In contrast, for experienced athletes—whose abilities may be approaching their performance ceiling—SVT may serve primarily to maintain and refine existing skills^62–64.

Limitations and future research directions

While the present findings provide preliminary support for the effectiveness of stroboscopic training, they should be interpreted with caution due to several limitations. First, the number of included studies was relatively small (n = 13), and participants were drawn from a range of sports disciplines. This variability may reduce the precision and generalizability of the meta-analytic estimates. Notably, certain subgroup analyses—such as those assessing training duration—were based on a limited number of studies, thereby increasing uncertainty in the corresponding effect size estimates. Second, although subgroup analyses accounted for some of the observed heterogeneity, unmeasured factors may still have influenced the outcomes. For example, the degree to which athletes depend on visual information may vary considerably across sports (e.g., shooting likely demands more intensive visual processing than endurance running), potentially affecting the efficacy of SVT. Third, many of the included trials had small sample sizes, resulting in limited statistical power, and few studies provided follow-up assessments. Most outcomes were measured only at the conclusion of the intervention, offering limited insight into the long-term retention of SVT effects. Consequently, it remains unclear how durable these performance gains are over time. Finally, the extent to which SVT translates into actual competitive success remains an open question. Current evidence is largely derived from laboratory-based or controlled training settings, using outcome measures such as reaction time and technical accuracy. Whether these improvements translate into tangible gains—such as increased scoring accuracy or faster decision-making during real competition—requires further empirical validation.

Future research should further investigate how stroboscopic training can be optimized and where its boundaries of application lie. A key direction is to examine whether SVT difficulty levels can be individualized based on athlete-specific characteristics. Subgroup analyses from the present study suggest that training parameters are not universally effective—athletes at different performance levels may respond better to stroboscopic stimulation strategies tailored to their unique needs and capacities. Additionally, future studies could explore the potential benefits of combining SVT with other perceptual training approaches. For example, integrating stroboscopic eyewear with target-tracking tasks or cognitive decision-making exercises may facilitate improvements across multiple perceptual–motor domains, thereby enhancing training efficacy through synergistic effects. Beyond the scope of competitive athletics, the broader applicability of SVT warrants further exploration. In rehabilitation settings, SVT may support the recovery of visual–cognitive function in individuals recovering from concussion. In military or driver training contexts, it may enhance reaction speed and decision-making under high-pressure conditions. As research on the clinical and applied potential of SVT continues to grow, such findings can inform practical implementation and contribute to a more comprehensive theoretical understanding of its underlying mechanisms. Moreover, as all studies included in this review were published in English, relevant findings from research published in other languages may have been overlooked. This language restriction may have limited the comprehensiveness and generalizability of the conclusions. Future reviews should consider expanding the scope of literature sources to include high-quality studies published in multiple languages, in order to provide a more balanced and objective assessment of the effects of SVT on collegiate athletes’ performance. Although this review employed multiple search strategies to ensure comprehensive literature coverage, some relevant studies may have been missed due to technical constraints and limited access to certain databases. This limitation may have affected the interpretive validity and generalizability of the findings. Future studies should aim to further optimize search procedures and strategies to enhance the systematicity and sensitivity of the retrieval process. Expanding the search scope and improving methodological rigor will help facilitate more robust and reliable quantitative evaluations of the effects of SVT on collegiate athletes’ performance based on larger and more representative datasets.

Conclusions

This systematic review and meta-analysis provide the first robust quantitative evidence supporting stroboscopic training as an effective intervention for enhancing sport-specific performance in athletes. The meta-analytic results yielded three key findings. First, SVT produced significant moderate effects on both time-based outcomes (SMD = −0.59, p < 0.01) and accuracy-based outcomes (SMD = 0.73, p < 0.01). Second, an effective training protocol was identified, comprising 6–10 weeks of training, 2–3 sessions per week, each lasting 10–20 min, with a strobe frequency of 5–20 Hz. Notably, optimal duty cycles appeared to vary by performance domain: 50–70% for time-based tasks and 10–50% for accuracy-based tasks. Third, the effectiveness of SVT was moderated by athlete characteristics. Adolescents (aged 10–18 years) exhibited greater improvements in time-based performance, while athletes with 4–8 years of training experience demonstrated the most pronounced overall gains. Collectively, these findings provide preliminary, evidence-based recommendations for coaches and practitioners seeking to implement SVT in applied sport settings. Nonetheless, future research should examine potential interactions among training parameters and investigate the neurophysiological mechanisms underlying SVT-induced performance enhancements across various sports and levels of expertise.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

Yuanyuan Luo contributed to the conceptualization, literature search, data curation, formal analysis, and writing of the original draft. Yingying Cao was responsible for the methodology, validation, interpretation of results, writing – review and editing, and project administration. Xinliang Pan contributed to methodology refinement, investigation and visualization, software support, and writing – review and editing. Shuairan Li performed formal analysis, software programming and data processing, visualization, and writing – review and editing. Denise Koh provided funding acquisition, resources, supervision, writing – review and editing, and served as the corresponding author. Yan Shi contributed to data collection and management, resource coordination, validation, and writing – review and editing. All authors have read and approved the final manuscript and share responsibility for the integrity of the work.

Data availability

Data Availability Statement All data analyzed in this systematic review and meta-analysis were obtained from publicly accessible peer-reviewed articles included in the final analysis. The datasets generated or analyzed during the current study are available from the corresponding author upon reasonable request.

Declarations

Competing interests

The authors declare no competing interests.

Ethical approval

As this study is a meta-analysis of previously published research, ethical approval was not required.