Introduction

Smartphone use has become a defining feature of modern life, shaping both personal and professional activities. By the end of 2021, the number of active smartphone users worldwide had reached 6.3 billion, with projections suggesting an increase to 7.34 billion by 20251. Alongside this rapid growth, research has documented a steady rise in the time devoted to daily tasks and social interactions through smartphones on a global scale2,3,4. While smartphones offer substantial conveniences and benefits, they also present persistent challenges, particularly in terms of health and well-being, that are increasingly recognized as significant concerns.

Prolonged smartphone use is associated with various adverse effects, most notably discomfort and fatigue in the neck and shoulders5,6,7,8,9. These problems are often attributed to the downward gaze posture, commonly referred to as “text neck syndrome” 10,11. Numerous studies have examined the musculoskeletal strain linked to forward head and neck flexion12,13,14,15,16, but eyestrain represents an additional concern. Prior work has largely approached eyestrain from epidemiological and optical perspectives17,18,19,20, leaving its ergonomic dimension underexplored. Although some investigations have incorporated gaze angle and viewing distance along with head and neck flexion14,15,21,22, the lack of direct, objective measures of visual fatigue has left the assessment of eyestrain during smartphone use largely speculative.

Critical flicker fusion frequency (CFF) refers to the threshold at which a flickering stimulus is perceived as continuous, reflecting the dynamic interaction between the visual system and the brain23. Although its sensitivity as a measure of visual load has been questioned in recent studies24,25, CFF remains widely regarded as a reliable indicator of eyestrain and visual fatigue26,27,28,29. In the context of smartphone use, reductions in CFF have been shown to signal increased visual and mental fatigue23. Beyond flicker-based measures, eye-movement parameters have also been applied to evaluate cognitive effort and visual efficiency in digital interaction contexts. For example, analyses of gaze behavior have demonstrated advantages of clean-code techniques and revealed characteristic patterns during unconventional software-tool usage30,31. These studies highlight the broader utility of ocular metrics in assessing attentional demands and visual strain across diverse human–computer interfaces.

Gaze angle, defined as the horizontal angle at which a user’s line of sight intersects the smartphone14,15,16,17,18,19,20,21, provides insight into the relationship between head–neck posture and device positioning. A larger gaze angle reflects a more pronounced downward gaze and has also been associated with “head phubbing,” or focusing on the smartphone while disengaging from social interaction32. Prior studies have shown that smartphone use while walking results in a greater downward gaze compared with stationary use2. Sarraf and Varmazyar15 further noted that the least favorable gaze angle occurs during standing, suggesting compensatory adjustments between neck flexion and gaze angle across postures. To maintain a healthy head–neck alignment, a gaze angle between 40° and 60° is typically recommended21. Another critical factor is viewing distance—the space between the eyes and the screen. Evidence consistently demonstrates that shorter viewing distances increase accommodative and vergence demands, thereby exacerbating eyestrain symptoms33,34,35. Whether these effects differ across smartphone usage modes, however, remains unclear.

Field observations of smartphone use have shown that factors such as posture (sitting, standing, or walking), hand usage (one- or two-handed), and participant characteristics (with or without musculoskeletal symptoms) significantly affect head and neck flexion angles10,14,15,36. At the same time, smartphones have become central tools for audio-visual engagement, with mobile gaming emerging as one of the most prevalent activities, particularly among youth37. Unlike browsing or texting, gaming often requires uninterrupted attention and sustained visual focus until specific objectives are achieved. This continuous interaction distinguishes gaming from other forms of smartphone use and may influence not only postural behavior but also visual load, potentially leading to different eyestrain outcomes than those reported in previous studies.

Building on these considerations, the present study investigated the impact of smartphone gaming on visual fatigue through a controlled comparison of two gaming modes (standing vs. walking) and two durations (15 vs. 30 min). Thirty participants (15 men and 15 women) completed simulated gaming sessions while measures were collected for CFF reduction, VFS, gaze angle, and viewing distance. We hypothesized that gaming while walking would induce greater eyestrain than standing, due to the additional demands on accommodation and vergence during dynamic movement. We further expected that prolonged gaming (30 min vs. 15 min) would exacerbate visual fatigue across both modes.

Materials and methods

Digital eyestrain can be categorized into external and internal symptoms, as described by Sheedy et al.36. Internal symptoms include blurred vision, diplopia, fatigue, and headaches, which arise from strain on the eye’s refractive and binocular systems. External symptoms are primarily associated with dry eye syndrome. From an ergonomic perspective, this study focused on internal symptoms by examining visual load during smartphone gaming in both standing and walking modes. To capture these responses, we employed four measures: reduction in CFF, VFS, gaze angle, and viewing distance.

We acknowledge that gaze angle and viewing distance occupy a dual role in our conceptual framework. While these variables represent immediate ergonomic responses to different gaming conditions, they may also function as intermediate factors that influence visual fatigue outcomes (CFF reduction and VFS). Based on established ergonomic models, we propose that: (1) Gaming mode and duration directly affect visual fatigue through cognitive load and visual attention demands; (2) Gaming mode and duration influence postural adaptations (gaze angle and viewing distance), which in turn may affect visual fatigue through altered accommodation and vergence demands. Given this dual nature, we analyzed these variables both as parallel outcomes and explored their potential mediating relationships through supplementary analyses.

Participants

A priori sample size estimation was performed using G*Power (Version 3.1.9.7) for a mixed-factorial analysis of variance (ANOVA), with sex as the between-subject factor and gaming mode and duration as within-subject factors. Assuming α = 0.05, power = 0.80, a large effect size (f = 0.45), and a correlation of 0.5 among repeated measures, the minimum required sample size was 28. The large effect size assumption was considered reasonable given the contrasting nature of standing versus walking conditions and the continuous, uninterrupted attention demands of gaming compared to typical smartphone use. To meet this requirement and ensure balanced groups, we recruited 30 healthy young adults (15 men and 15 women) from the university community, consistent with sample sizes reported in previous smartphone ergonomics studies2,10,26.

All participants reported at least three hours of daily smartphone use for more than one year, had normal or corrected-to-normal visual acuity (20/20 or better with correction if needed), and none had a history of musculoskeletal disorders or visual impairments. Male participants had a mean ± SD age of 21.4 ± 1.3 years, height of 174.2 ± 5.6 cm, and body mass of 72.0 ± 10.4 kg, whereas female participants averaged 22.0 ± 1.1 years, 159.3 ± 4.6 cm, and 50.7 ± 7.4 kg. Baseline CFF values averaged 38.3 Hz for both sexes, within the normal adult range of 35–40 Hz23 (Table 1). Baseline CFF values were consistent across all conditions and sexes, with mean differences below 0.3 Hz and within-participant variability around ± 2 Hz, confirming that pre-task visual states were stable before each experimental session. Males were slightly younger and significantly taller (p < 0.001) and heavier (p < 0.001) than females. All participants received a detailed explanation of the study protocol and provided written informed consent before participation.

Table 1 Basic information of male and female participants in the study.
Full size table

CFF measurement

CFF was employed as an objective indicator of visual fatigue because it is non-invasive, demonstrates high test–retest reliability, and is sensitive to visual performance changes induced by prolonged near work. Although its sensitivity under specific visual tasks has been debated, CFF remains widely applied in ergonomics and ophthalmology research, including smartphone-related fatigue assessments39. In this study, a reduction in CFF was interpreted as increased visual fatigue40,41,42.

CFF was measured using a Handy Flicker device (HF-II, Neitz, Tokyo, Japan). For each trial, ascending and descending thresholds were recorded and averaged to determine pre- and post-task CFF. Each measurement was repeated three times, and the mean value was used for analysis. All measurements were conducted under consistent indoor lighting (~ 500 lx) at the same time of day, with ascending–descending sequences performed in a fixed order for all participants. In the ascending procedure, flicker frequency increased from 20 Hz until the stimulus was perceived as stable, whereas in the descending procedure, frequency was gradually reduced until flicker was again detected, following the method of Gautam and Vinay23. The average of the ascending and descending thresholds represented the participant’s CFF for that trial.

Subjective visual fatigue rating

Subjective visual fatigue was assessed using the questionnaire developed by Heuer et al.43, which has been widely applied in visual fatigue research44,45,46,47,48. The questionnaire includes six items: (1) difficulty seeing, (2) strange sensations around the eyes, (3) tired eyes, (4) numbness, (5) dizziness when viewing the screen, and (6) headache. Each item was rated on a 10-point scale, where 1 indicated “not at all” and 10 indicated “extremely serious.” The mean score across the six items was calculated as the VFS. To account for baseline differences, the change in VFS was analyzed by subtracting the pre-task score from the post-task score for each gaming session.

Gaze angle and viewing distance measurements

Gaze angle and viewing distance were determined from symmetrical sagittal photographs of each participant (Fig. 1). Digital markings were performed using CorelDRAW (Graphics Suite 2023 v24.5, Corel Co., Ottawa, ON, Canada). The gaze angle was defined as the line between the eyeball (E) and the midpoint of the smartphone’s length (M), while viewing distance was measured as the spatial separation between these two points. A scale fixed to the wall was used to calibrate viewing distance measurements. The reliability of this method has been validated by Chen et al.14, who reported measurement errors of less than 3° for angles and 1.3 mm for distances, confirming its precision and reproducibility.

Fig. 1
figure 1

Diagram depicting marker placement and definitions of gaze angle (GA) and viewing distance (VD) for the human body, along with operational postures (left) and a screenshot from the study’s adopted game (right).

Full size image

Experimental design and procedure

Each participant completed four smartphone gaming trials, combining two gaming modes (standing and walking) with two gameplay durations (15 and 30 min). This 2 × 2 design yielded four testing conditions performed in a randomized order to minimize carryover effects. CFF and VFS were measured at the beginning and end of each session, while gaze angle and viewing distance were recorded during the final 1-min interval at 15-s intervals and averaged for analysis. This end-of-trial sampling strategy was adopted to minimize the influence of early transient adjustments and to capture participants’ stabilized viewing posture after full task engagement. Pilot observations indicated that both gaze and viewing distance typically plateaued within the first few minutes of gameplay, consistent with previous findings that smartphone-related postural adaptations reach steady state during sustained use9,22. Therefore, sampling during the last minute was deemed representative of the steady-state posture rather than an end-of-task settling effect. The puzzle game TWENTY (developed by Stephen French; Fig. 1) was used for all trials.

Participants used their own smartphones to enhance ecological validity, with the game pre-installed prior to testing. Screen sizes ranged from 5.5 to 6.5 inches with brightness set according to individual preference. Although device variability could influence absolute visual demands, the within-subject design controlled for these differences. In the walking conditions, participants walked on a treadmill (CS-5728, Chanson, Taipei) at their preferred, self-selected pace, ensuring a natural walking pattern representative of daily mobility. In the standing condition, they remained stationary on the same treadmill, which was set to 0 speed. To reduce fatigue and experimental error, at least 10 min of rest was provided between trials. During rest periods, participants were free to engage in non-visual activities but were instructed to avoid using smartphones or other digital screens.

Statistical analysis

Data were analyzed using SPSS 23.0 (IBM Corp., Armonk, NY, USA) and R 4.3.0 (R Core Team, 2023) with the lme4 package. Primary analyses employed three-way ANOVA to examine the effects of sex (between-subject: men, women), gaming mode (within-subject: standing, walking), and gameplay duration (within-subject: 15, 30 min) on all four dependent variables. CFF reduction was designated as the primary outcome variable because it provides an objective index of visual fatigue, whereas VFS, gaze angle, and viewing distance were treated as secondary supportive measures. When significant effects were identified, post-hoc comparisons were performed using t tests.

To address multiple testing across the four related outcomes, a Bonferroni correction (α = 0.0125) was applied when interpreting omnibus F tests. Because both within-subject factors contained only two levels, the assumption of sphericity was inherently met; Mauchly’s test was checked for completeness and confirmed this. Prior to analysis, the Shapiro–Wilk test verified that the distributions of all variables did not deviate from normality (p > 0.05), and Levene’s test confirmed homogeneity of variances (p > 0.05), validating the use of ANOVA. Partial η2 was calculated for each omnibus effect to indicate the proportion of variance explained. Effect sizes for pairwise comparisons were expressed as Cohen’s d with 95% confidence intervals to illustrate the magnitude and precision of observed effects49. Effects meeting the corrected threshold (p < 0.0125) are reported as “significant,” while those with 0.0125 < p < 0.05 are described as “trends”.

To complement the ANOVA framework, linear mixed-effects models (LMMs) were also considered. Because each participant completed all four conditions and the design was fully balanced, repeated-measures ANOVA and LMMs with random intercepts yield equivalent fixed-effect estimates. Given the modest sample size (n = 30) and the limited number of within-subject conditions, ANOVA was retained for clarity and interpretability, while acknowledging that LMMs could flexibly incorporate random slopes and covariates such as baseline CFF, participant height, or device characteristics in larger samples. In future studies, mixed-effects modeling will be adopted to account for individual heterogeneity more fully.

To address the potential mediating role of postural variables, we conducted supplementary mediation analyses using linear mixed-effects models to account for repeated measures. The mediation analysis followed the approach of Baron and Kenny50 adapted for multilevel data, testing whether viewing distance and gaze angle mediate the relationship between gaming mode and visual fatigue outcomes (CFF reduction and VFS). The percentage of mediation was calculated as (indirect effect/total effect) × 100. Statistical significance was set at p < 0.05 for all analyses.

Ethics approval

This study was approved by the Ethics Committee of National Taiwan University, Taiwan (protocol code NTU-REC 202312-EM-051) and conducted in accordance with the Declaration of Helsinki.

Results

Table 2 summarizes the three-way ANOVA outcomes for the effects of gaming mode, gameplay duration, and sex on visual-fatigue and postural responses. Standardized effect sizes are presented as partial η2 for omnibus effects and Cohen’s d with 95% confidence intervals for two-level contrasts (sex, mode, duration). Cohen’s d was not computed for higher-order interactions because such effects represent moderation rather than a single pairwise mean difference. As shown in Table 2, sex significantly affected gaze angle and viewing distance (both p < 0.001), with very large between-sex differences (Cohen’s d = 2.05–2.11, 95% CI≈1.3–2.9). Gaming mode (walking vs. standing) exerted significant effects on all measures (p ≤ 0.002), with large within-subject effects (Cohen’s d = 0.63–0.83; ηₚ2 = 0.29–0.42). Longer gameplay duration (30 min vs. 15 min) also increased CFF reduction and VFS (p < 0.001; d ≈ 0.7–0.8; ηₚ2 = 0.39–0.41). The duration effect on viewing distance approached significance (p = 0.046) but did not survive Bonferroni correction (α = 0.0125). No significant interactions among sex, mode, and duration were observed.

Table 2 Three-way ANOVA results for the effects of gaming mode (standing vs walking), gameplay duration (15 vs 30 min), and sex on visual-fatigue and postural measures.
Full size table

The paired comparisons of these responses across participant sex, gaming mode, and gameplay duration are presented in Figs. 3, 4, and 5, respectively, offering a visual depiction of the main and interaction effects that complements the statistical findings and highlights the consistent directional patterns observed across conditions.

To further clarify the underlying mechanisms linking gaming mode to visual fatigue, we conducted mediation analyses to test whether posture-related factors—specifically gaze angle and viewing distance—served as intermediaries between gaming mode and the visual fatigue outcomes (CFF reduction and VFS). Table 3 summarizes the mediation results based on linear mixed-effects models. These analyses allowed us to determine whether postural adaptations contributed directly to increased visual strain or merely accompanied changes in task condition.

Table 3 Mediation analysis of posture-related variables (viewing distance, gaze angle) on the effect of gaming mode (standing vs walking) on visual-fatigue outcomes (CFF reduction and VFS). Linear mixed-effects models were used to estimate total, direct, and indirect effects.
Full size table

For CFF reduction, the total effect of gaming mode (walking vs. standing) was β = 1.20 (p < 0.001). When viewing distance was included as a mediator, the direct effect decreased to β = 0.89 (p < 0.001), and the indirect effect through viewing distance was β = 0.31 (p < 0.05), corresponding to 26% mediation. Similarly, for VFS, the total effect of β = 1.10 (p < 0.001) was reduced to a direct effect of β = 0.82 (p < 0.001) after controlling for viewing distance, with an indirect effect of β = 0.28 (p < 0.05), representing 25% mediation. Gaze angle showed weaker mediating effects, accounting for 12% of the gaming-mode effect on CFF reduction (p = 0.082) and 13% on VFS (p = 0.121), neither of which reached statistical significance. These findings suggest that viewing distance partially mediates the relationship between gaming mode and visual fatigue, whereas gaze angle contributes only marginally.

Discussion

With the rapid growth of mobile gaming, understanding how different gaming modes and exposure durations influence eyestrain is increasingly important. In this controlled study of young adults, gaming while walking produced greater objective (CFF reduction) and subjective (VFS) visual fatigue than standing, and longer sessions (30 min vs. 15 min) further intensified these effects. Sex differences were observed for posture-related variables (gaze angle and viewing distance) but not for fatigue measures, indicating that anthropometric and postural characteristics differ by sex, whereas visual fatigue responses are primarily determined by gaming mode and duration. Collectively, these findings demonstrate that dynamic locomotion compounds visual load beyond static use, emphasizing the heightened visual strain risk during mobile gaming and prolonged, uninterrupted play.

Gaming mode effect on the responses

Previous studies have examined smartphone use across different postures such as sitting, standing, and walking2,10,12,14,15,51,52,53, but the specific context of gaming has been largely overlooked. Most work has focused on browsing or texting, often highlighting variations in neck flexion and muscle activity. For example, Yoon et al.10 reported greater neck–shoulder muscle activation during walking than standing, with reduced neck flexion in dynamic conditions. Consistent with these findings, our results showed smaller gaze angles during gaming while walking compared with standing (Table 2, Fig. 2), suggesting that environmental demands during walking may limit head tilting and promote more downward gaze behavior.

Fig. 2
figure 2

Comparisons of main effects across sex, gaming mode, and gameplay duration variables for all responses (Error bars represent + 1 standard deviation of the mean).

Full size image

The greater visual fatigue during walking may be partially explained by challenges to the vestibulo-ocular reflex (VOR). During normal walking, the VOR stabilizes retinal images through compensatory eye movements. However, when viewing a smartphone while walking, the screen moves with the head, requiring VOR suppression—a demanding process that can cause retinal image blur and contribute to visual fatigue. Additionally, head oscillations during walking (5–8 cm vertically at typical speeds) create dynamic viewing conditions that further challenge visual stability54. The self-selected walking speeds in our study varied from 3.2 to 4.8 km/h (mean 4.0 ± 0.4 km/h), and faster walking speeds would produce greater head movements and more challenging VOR suppression demands, potentially explaining some of the individual variability in visual fatigue responses. Future studies could benefit from direct measurement of head movements and VOR gain during smartphone walking.

Importantly, gaming mode significantly affected both subjective and objective indicators of visual fatigue. Walking produced greater CFF reduction (2.8 Hz vs. 1.6 Hz) and higher VFS scores (4.5 vs. 3.4) than standing, indicating intensified eyestrain during dynamic movement. This aligns with evidence that cervical spine motion during walking can strain ocular accommodation and vergence mechanisms10,12,55. Neurophysiological studies further support this interpretation; Lee et al.56 found that visual presentations deviating from natural viewing conditions altered brainwave activity and increased visual fatigue, paralleling the elevated strain observed when gaming while walking. The shorter viewing distances we observed in walking mode (29.7 cm vs. 33.3 cm; Fig. 2) may also reflect this heightened visual load.

Few studies have examined ocular strain during walking. Leung et al.2 reported that smartphone use while walking altered corneal astigmatism and induced greater esophoric near heterophoria compared with sitting, underscoring key differences between dynamic and static viewing conditions. While their study did not include a standing condition, our findings extend this evidence by showing that mobile gaming while walking elicits both greater visual fatigue and altered postural responses relative to standing.

In this study, smaller gaze angles and shorter viewing distances were observed when gaming while walking, yet these changes do not necessarily reduce eyestrain. A smaller gaze angle may slightly decrease ocular surface exposure and dryness18, but the concurrent reduction in viewing distance increases accommodative and vergence demands, heightening internal visual strain33,34,35. Moreover, prolonged downward gaze can suppress blinking and disrupt tear dynamics, further aggravating fatigue20. Hence, while display position influences viewing geometry, overall eyestrain reflects the trade-off between surface exposure and accommodation load. In the present study, the shorter viewing distance during walking likely outweighed any potential benefit from a smaller gaze angle, as evidenced by greater CFF reduction and higher VFS. This interpretation aligns with ergonomic findings that shorter working distances are a major contributor to digital eyestrain43,44.

Display characteristics also contribute importantly to visual fatigue. Prior studies show that screen factors such as resolution, curvature, color mode, and luminance contrast can influence comfort and performance during sustained viewing13,14,15,16,17. Curved or high-resolution displays may reduce distortion and improve legibility, whereas excessive brightness or glare can heighten accommodative stress and tear-film instability, leading to fatigue18,19. These findings suggest that visual strain results not only from user posture and exposure duration but also from interactions between display properties and ambient lighting.

Sex-based difference

Sex did not significantly affect objective (CFF reduction) or subjective (VFS) visual fatigue, but clear differences appeared in posture-related variables: women showed smaller gaze angles and shorter viewing distances than men. These differences may reflect a combination of anthropometry and postural strategy, but are not fully explained by body size alone. For instance, Brink et al. reported no strong correlation between height and body-angle measurements within sex, suggesting that stature is not the sole determinant of posture during device use57. Nevertheless, height and arm length could influence the natural viewing geometry; incorporating these covariates in future analyses would help determine whether the observed sex-related posture differences persist after adjusting for body size58.

Regarding viewing distance, sex differences likely reflect handling habits and visual strategies during phone use rather than size-related constraints; prior work documents variability in typical smartphone viewing distances across users and contexts59. Although participants were instructed to hold the phone naturally, we did not systematically record grip strategy or hand usage. Differences such as one- versus two-handed operation or thumb reach could influence both gaze angle and viewing distance, potentially accounting for some between-sex variation. Future research should include direct observation or video-based classification of hand posture to clarify these behavioral factors.

Because shorter viewing distances increase accommodative and vergence demands, they are associated with greater eyestrain risk. In our data, however, these postural differences did not translate into sex differences in visual fatigue, indicating that task demands (mode and duration) may be the primary drivers of fatigue. Given the modest sample size (15 participants per sex) and the exploratory nature of this comparison, the findings should be interpreted cautiously. Larger samples with explicit control of anthropometric and behavioral covariates are needed to determine whether posture differences reflect intrinsic sex effects or contextual strategies.

Gameplay time effect

Extended gaming sessions (30 min) significantly increased both subjective (VFS) and objective (CFF) indicators of visual fatigue across sexes (Fig. 3). This outcome aligns with evidence that prolonged near-vision tasks exacerbate fatigue through sustained accommodation, convergence demands, and exposure to screen glare or flicker26,60,61.

Fig. 3
figure 3

Comparisons of four measured responses between sexes across two gaming durations (Error bars represent ± 1 standard deviation of the mean).

Full size image

Beyond fatigue measures, viewing distance showed a nonsignificant trend toward reduction with longer duration (p = 0.046). Descriptively, this pattern appeared more pronounced during walking (3.8 cm reduction) than standing (1.5 cm reduction), and appeared stronger among female participants (Fig. 4), though the interaction was not statistically significant. These patterns suggest possible duration- and sex-specific postural adaptations during prolonged smartphone use that warrant confirmation in larger samples. Prior work by Long et al.35 also documented progressive viewing distance reductions over a 1-h session, although their small, mixed-sex sample precluded sex-specific insights.

Fig. 4
figure 4

Comparisons of four measured responses between sexes for two gaming modes (Error bars represent ± 1 standard deviation of the mean).

Full size image

Interestingly, gaze angle showed descriptively different patterns across modes, though the Mode × Duration interaction was not statistically significant (p = 0.129). Gaze angle remained stable during walking (56.0° at 15 min vs. 56.8° at 30 min), whereas standing sessions showed a numerical increase (61.2° vs. 65.7°). This pattern suggests that extended gaming in a static position may encourage a greater downward gaze, potentially heightening the prevalence of phubbing-like head posture, though this requires confirmation in larger samples. Notably, gaze angles during standing exceeded the recommended ergonomic range of 40–60° 21.

The observed reduction in viewing distance during longer play, particularly while walking, reinforces the mediating role of visual–postural factors in cumulative fatigue. Consistent with the mediation results (Table 3), shorter viewing distances increase accommodative and vergence demands, thereby intensifying both objective (CFF) and subjective (VFS) fatigue measures. This pattern suggests that prolonged exposure exacerbates visual strain partly through progressive reductions in viewing distance, linking task duration to eyestrain via an accommodation-driven pathway (Fig. 5).

Fig. 5
figure 5

Comparisons of four measured responses between gaming durations within two modes (Error bars represent ± 1 standard deviation of the mean).

Full size image

Mediation analysis

The mediation analysis provides important insights into the causal pathways linking gaming mode to visual fatigue. Viewing distance emerged as a significant partial mediator, accounting for approximately 25–26% of the effect of gaming mode on both CFF reduction and VFS. This finding supports the accommodation-vergence theory, whereby shorter viewing distances during walking increase accommodative demand and convergence stress, thereby contributing to visual fatigue. However, the substantial direct effects remaining (approximately 74–75%) indicate that additional mechanisms—such as visual-vestibular conflicts during walking, divided attention, and dynamic visual demands—play major roles in gaming-induced eyestrain. The marginal mediating effect of gaze angle suggests that while downward gaze may contribute to fatigue, its role is secondary to viewing distance.

Practical implications

The findings of this study highlight several applied considerations. The observed sex-related differences in gaze angle and viewing distance suggest that game developers should design interfaces that account for anthropometric variation. Features such as scalable font sizes, adjustable contrast, and flexible layouts may reduce visual strain across diverse user groups62. In addition, the greater fatigue observed while walking compared with standing underscores the need for user education and context-aware reminders that discourage visually demanding tasks during motion.

Prolonged sessions (30 min) produced more pronounced reductions in CFF and increases in VFS, emphasizing the value of timed rest breaks, in-game reminders for eye exercises, and adjustable viewing distance prompts to mitigate accommodative and vergence strain4. Raising awareness among gamers about these risks can further encourage healthier device use, while eyecare professionals should remain attentive to potential optical aftereffects and vergence adaptations following extended smartphone engagement, as also noted by Leung et al.2.

Study limitations

This study has several limitations. First, the sample size of 30 participants (15 men, 15 women), though determined by a priori power analysis, primarily supports within-subject effects of gaming mode and duration; therefore, sex-related differences should be regarded as exploratory. The sample consisted solely of healthy young adults, limiting generalizability to other age groups or individuals with visual or musculoskeletal disorders. Participants used their own smartphones to enhance ecological validity, but variability in device size, weight, and display characteristics may have introduced uncontrolled differences in visual and postural demands. The tested durations (15 and 30 min) were shorter than typical gaming sessions, and laboratory lighting differed from real-world conditions involving glare, reflections, or motion. Moreover, only one casual puzzle game—chosen for its moderate cognitive and visual demands—was examined, so findings should not be generalized to high-motion or text-dense games with greater oculomotor or attentional loads.

Head movement and vestibulo-ocular reflex (VOR) function were not directly measured, although both may contribute to visual fatigue through retinal image blur during walking. Standardizing walking speed in future studies could reduce variability from head motion. Certain measurement and environmental factors could also be refined to strengthen internal validity. Although CFF testing was conducted under controlled indoor lighting, ambient illumination, time of day, and the fixed order of ascending–descending sequences were not randomized, which may subtly influence flicker thresholds40,42. Because all four sessions occurred within a single visit, learning or fatigue effects cannot be excluded; incorporating session order as a covariate would help distinguish true exposure-related changes. Allowing participants to use preferred screen brightness enhanced ecological validity but introduced variability in retinal illuminance. Finally, posture was assessed only during the final minute of each trial—providing stable steady-state data but potentially overlooking short-term fluctuations earlier in gameplay. Continuous optical tracking could better capture such dynamic postural adjustments in future work.

Finally, mediation modeling was limited by sample size and concurrent measurement of posture and fatigue. Larger, multi-day field studies incorporating unobtrusive viewing-distance tracking, blink or tear-film measures, and ecological momentary assessment would extend these laboratory findings and guide ergonomic design for mobile gaming.

Conclusions

Mobile gaming while walking has become increasingly common in daily life. This study demonstrated that playing smartphone games while walking induces greater objective and subjective visual fatigue than gaming in a stationary standing posture. Extending gameplay from 15 to 30 min further amplified fatigue, underscoring the role of duration in visual strain. Additionally, exploratory analyses revealed sex-based differences in posture, with women showing closer viewing distances and smaller gaze angles compared to men, though these did not translate to differences in visual fatigue outcomes.

These findings not only provide evidence-based insights for game developers to optimize interface design and reduce visual demands but also highlight the importance of raising user awareness about the risks of prolonged smartphone gaming. Encouraging healthier postural habits and moderated use may help mitigate the potential adverse effects on visual comfort and well-being.