Evaluation of viewing distance on vision tasks using virtual reality technology for children with intermittent exotropia

Abstract

Purpose:

This study sought to investigate dynamic visual acuity (DVA), binocular functions, and perceptual eye position (PEP) measured via a virtual reality (VR) evaluation system in children with intermittent exotropia (IXT) at varying viewing distances.

Design:

Retrospective, case-control observational study.

Methods:

A retrospective study was conducted involving 118 children: 59 with IXT and 59 age-matched controls. Comprehensive ophthalmic assessments, including DVA, static and dynamic PEP, contour integration, and stereopsis, were performed using an intelligent VR-based visual perception platform. Testing was conducted at four viewing distances: 0.4 m, 0.7 m, 1 m, and 3 m.

Results:

Children with IXT demonstrated significantly higher rates of monocular DVA abnormalities across all distances compared to controls (all P < 0.001). Abnormal contour integration was markedly worse in the IXT group at 1 m (52.54% vs. 28.81%; P = 0.007). Both groups exhibited the poorest DVA and contour integration at the 3 m distance. Receiver operating characteristic (ROC) curve analysis identified optimal PEP thresholds for contour integration assessment in IXT and control groups, with high sensitivity and specificity.

Conclusions:

VR-based testing offers a novel, sensitive tool for detecting visuomotor deficits in children with IXT, particularly at critical distances. This methodology may enhance early diagnosis and individualized treatment planning for visual function deficits in pediatric populations.

Intermittent exotropia (IXT) is one of the most common forms of childhood strabismus.[1,2] Conventional clinical evaluations of IXT primarily focus on static parameters such as visual acuity and angle of deviation, which are typically assessed using the prism cover test under fixed visual conditions.[3,4] However, these static assessments often fail to capture the dynamic nature of visual function in real-world settings, where gaze continuously shifts to sample visual information from the environment. Unlike static eye alignment, dynamic visual function is highly dependent on factors such as gaze position, target distance, and context. Emerging evidence suggests that the control of ocular alignment in IXT can vary significantly over the course of a day, influenced by viewing distance and other situational demands.[5,6] This variability in control is particularly critical, as it impacts essential visuomotor tasks, including reading, obstacle navigation, and interaction with digital screens. Prior studies have also identified deficits in contour integration and stereopsis in IXT, even in the absence of reduced visual acuity, highlighting the multifaceted nature of the disorder.[7,8]

Given the increasing integration of digital devices into daily life, evaluating visual function across a range of distances is clinically relevant. Different viewing distances correspond to distinct visual tasks: 0.4 m reflects activities such as reading and smartphone use, 0.7 m corresponds to computer work, 1 m involves tasks requiring fine hand-eye coordination, and 3 m represents interactions such as watching television or interpersonal communication. Understanding distance-specific visual performance is essential for developing targeted interventions and establishing thresholds for functional recovery in IXT.

Virtual reality (VR) technology offers a unique opportunity to evaluate binocular visual function in a dynamic, controlled environment that closely simulates real-world conditions.[9,10] By leveraging VR, it is possible to assess key visual parameters such as dynamic visual acuity (DVA), perceptual eye position (PEP), contour integration, and stereopsis across varying viewing distances. This study aims to bridge the gap between clinical testing and real-world visual demands by systematically exploring the impact of viewing distance on binocular visual function in children with IXT and healthy controls.

Methods

This retrospective study included 118 children aged 4 to 15 years, comprising 59 children diagnosed with IXT and 59 age-matched healthy controls with no history of eye disease other than refractive error. The study was approved by our Institutional Review Board and was conducted in strict adherence to the tenets of the Declaration of Helsinki (2024-436A01). Informed consent was obtained from all participants or their guardians before enrollment. Inclusion criteria for the IXT group required a distance deviation of ≥ 20 prism diopters (PD) on the prism cover test. Control participants exhibited orthotropia or exophoria and had binocular visual acuity of ≥ 20/40 (0.5 logMAR) or better. Refractive error, assessed through cycloplegic refraction, ranged from − 5.375 diopters to + 4.00 diopters. Exclusion criteria included amblyopia, high refractive error (≥6.00 diopters), structural ocular anomalies, a history of ocular surgery, or any neurological conditions.

VR-based visual function assessment

Assessments were performed using a VR-based intelligent platform designed to evaluate DVA, PEP, contour integration, and stereopsis at four distinct viewing distances (0.4 m, 0.7 m, 1 m, and 3 m). DVA: The DVA test utilized a dynamic optotype (Dyop) that rotates at a rapid pace of 450 milliseconds per revolution. The moving segmented optotype, designed to dynamically stimulate a 16-photoreceptor cluster with a minimum area of resolution (MAR) of 0.54 arc minute squared, assessed dynamic acuity. Participants wore red–blue dichoptic glasses and were tasked with discerning the direction of rotation of Dyops presented at four distances: 0.4 m, 0.7 m, 1 m, and 3 m. The direction of rotation changed randomly upon each appearance to maintain participant engagement. At the same distances, participants also evaluated the relative depth between a central cross and a Dyop positioned above it.

Contour integration and stereopsis

Testing procedure: When the eyes are positioned outside the circle, they perceive flashing points at various positions around the circle over time. The visual cortex integrates the information from both eyes to determine the correct rotation direction of the circle. At four distinct distances, the accuracy of judging the rotation direction of the lowermost circle is used to assess whether the contour integration ability is functioning normally. In the out-of-eye state, the topmost circle and the middle cross establish a front-back relationship through varying parallax. The depth relationship between the cross and the circle is compared at four distinct distances [Fig. 1].

Figure 1.

The “Dyop” visual acuity target. The observer is required to indicate which of the targets is spinning (right or left), and the direction of rotation (clockwise or counter-clockwise)

PEP

Participants equipped with red–blue glasses for dichoptic viewing, perceived the radiating cross and circular target separately with each eye. [Fig. 2]. Using a mouse, participants aligned the cross to the center of the circle. Horizontal and vertical deviations were automatically recorded by the system and converted into angular values (5 pixels = 0.1° angular deviation). A full 360° sequence of eye position displacement and correction was meticulously documented [Fig. 3].

Figure 2.

Stimulus design of perceptual eye position (PEP): The background luminance was 44 cd/m² with a gray hue spanning a visual angle of 38° ×18°. The circular target measured 0.8° ×0.8°, while the radiating cross spanned 0.66° ×0.66°. Luminance for white and black elements averaged 80 cd/m² and 30 cd/m², respectively

Figure 3.

Participants aligned the cross to the center of the circle by equipping with red–blue glasses for dichoptic viewing. Horizontal and vertical deviations were automatically recorded by the system and converted into angular values (5 pixels = 0.1° angular deviation)

Statistical analysis

Statistical analysis was conducted using commercially available software (SPSS version 16.0, SPSS, Chicago, IL). The normality of all data samples was assessed using the Kolmogorov-Smirnov test. A two-tailed paired student’s t-test was employed to compare data between the two scanning modes at each examination time for normally distributed data, whereas the Wilcoxon signed-rank test was used for non-normally distributed data. Chi-square tests were applied to analyze demographic data. The correlation between PEP and binocular vision functions tested using VR technology was evaluated through logistic regression analysis. Receiver operating characteristic (ROC) curve analysis was performed to determine the optimal cut-off values of PEP required for interpreting binocular visual perceptual abilities, including contour integration and stereopsis, at varying viewing distances. ROC analysis also provided sensitivity and specificity metrics for the parameters. A P value of less than 0.05 was considered statistically significant to ensure the validity of the findings.

Results

Demographics of patients

A total of 118 children participated in this study, with 59 diagnosed with IXT and 59 serving as healthy controls. The mean age of the IXT group was 8.25 ± 2.93 years (range: 4–15 years), while the control group had a mean age of 6.98 ± 2.15 years (range: 4–14 years). Significant differences were identified between the groups in uncorrected distance visual acuity (UDVA), spherical equivalent refractive error, and cylindrical power (all P < 0.01) [Table 1]. These differences underline the underlying visual characteristics unique to IXT patients.

Table 1.

Baseline demographic and clinical characteristics for the participants

Group	Intermittent exotropia (n=59)	Control (n=59)	P
Age (y)
Mean±SD	8.25±2.93	6.98±2.15	P=0.013
Range	4-15	4-14
Gender (F/M)	26/33	41/18
Visual acuity
LogMAR UDVA
Mean±SD	0.28±0.28	0.12±0.14	P<0.001
LogMAR CDVA
Mean±SD	0.09±0.10	0.10±0.11	P=0.733
Sphere
Mean±SD	−0.21±1.37	0.41±0.71	P<0.001
Range	−4.75-2.25	−1.25-3.25
Cylinder (mean±SD)
Mean±SD	−0.20±0.91	0.04±0.55	P=0.02
Range	−0.275-2.25	−1.25-2.25
SE (mean±SD)
Mean±SD	−0.35±1.58	0.43±0.84	P<0.001
Range	−5.375-2.5	−1.50-4.25
Angle of deviation (PD)
Distance	−26.02±9.67	NA
Near	−24.38±11.42	NA

Comparison of monocular and binocular DVA at different viewing distances

Across all viewing distances, the IXT group exhibited significantly higher rates of monocular DVA abnormalities compared to the control group (all P < 0.001). Notably, the highest abnormality rates were recorded at a viewing distance of 3 m, with 51.69% in the IXT group and 38.14% in the control group [Fig. 4a]. A progressive decline in performance was observed with increasing viewing distance in both groups, with the poorest performance consistently seen at 3 m [Fig. 4b].

Figure 4.

(a to d). Comparison of abnormal rates of DVA (a and b), stereopsis (c), and contour integration (d) between the two groups at varying distances. *P < 0.01

Comparison of contour integration at different viewing distances

At a viewing distance of 1 m, the IXT group demonstrated significantly poorer contour integration performance compared to the control group (52.54% vs. 28.81%; P = 0.007) [Fig. 4c]. This difference emphasizes the impact of IXT on intermediate-range visual processing. For both groups, performance in contour integration declined as the viewing distance increased, with the most pronounced deficits observed at 3 m, highlighting the role of distance in visual integration tasks [Fig. 5a and b].

Figure 5.

Abnormal rates of DVA, stereopsis, and contour integration from the IXT group (a) or control group (b) at varying distances. For both groups, the abnormal proportion of monocular and binocular DVA viewing at 3 m was significantly higher than that at 0.4 m, 0.7 m, and 1 m (all P < 0.01)

Comparison of stereopsis at different viewing distances

No siginificant difference was found in stereopsis between two groups at various distance [Fig. 4d]. In both groups, stereopsis performance decreased with increasing viewing distance, indicating a general difficulty in perceiving depth at longer distances [Fig. 5a and b]. These findings suggest that while stereopsis may remain intact in IXT patients, its effectiveness diminishes as distance increases.

Comparison of PEP for dynamic and static assessment

As shown in Table 2, PEP deviations, both horizontal and vertical, were significantly greater in the IXT group compared to the control group for both dynamic and static assessments (all P < 0.001). PEP deviations, both horizontal and vertical, were significantly greater in the IXT group compared to the control group for both dynamic and static assessments (all P < 0.001). ROC curve analysis identified optimal PEP thresholds for predicting contour integration deficits, demonstrating high sensitivity and specificity. Specifically, thresholds were determined to be 252.055 pixels for horizontal deviation and 8.90 pixels for vertical deviation at a viewing distance of 1 m. [Table 3A, Fig. 6a]. However, for the control group, thresholds were determined to be 6.07 pixels for vertical deviation at a viewing distance of 0.4 m and 4.46 pixels at a viewing distance of 1 m [Table 3B, Fig. 6b and 6c]. These thresholds provide valuable markers for detecting binocular visual function impairments in children with or without IXT.

Table 2.

Comparison of perceptual eye position (PEP) between two groups

Groups	Horizontal visual PEP pixels (static)	Vertical visual PEP pixels (static)	Horizontal visual PEP pixels (dynamic)	Vertical visual PEP pixels (dynamic)
Intermittent exotropia	273.20±259.00	36.63±41.62	293.07±285.45	30.76±40.30
Control	52.04±94.68	11.15±13.58	59.13±123.41	8.09±9.54
P	0.000	0.000	0.000	0.000

Table 3A.

ROC curve analysis for interpreting the results of the contour integration in the IXT group at viewing distance of 1 m

Parameter	AUC (%)	P	Threshold value (pixels)	Sensitivity/Specificity
Horizontal PEP (dynamic)	68.3	0.016	252.055	0.65/0.75
Vertical PEP (dynamic)	72.3	0.003	8.90	0.90/0.54

Figure 6.

ROC curves for perceptual eye position (PEP) deviations thresholds predicting visual function deficits. (a)Thresholds were determined to be 252.055 pixels and 8.90 pixels for PEP at a viewing distance of 1 m in IXT group. (b to c) PEP thresholds were determined to be 6.07 pixels for vertical deviation at a viewing distance of 0.4 m and 4.46 pixels at a viewing distance of 1 m in the control group

Table 3B.

ROC curve analysis for interpreting the results of the contour integration and stereopsis in the control group at viewing distances of 0.4 m and 1 m, respectively

Parameter	AUC (%)	P	Threshold value (pixels)	Sensitivity/Specificity
Vertical PEP (0.4 m)	77.7	0.004	6.07	0.85/0.65
Vertical PEP (1 m)	72.3	0.004	4.46	0.85/0.56

Discussion

Our findings revealed that the proportion of monocular DVA abnormalities in the IXT group was significantly higher than that of the control group across all viewing distances. This observation may be attributed to the physiological characteristics of eye movements and the biomechanical properties of IXT.[11,12] In individuals with normal binocular function, visual feedback facilitates accurate fixation on targets, maintaining near-perfect ocular alignment.[8] However, the inability of IXT to fuse images means that one eye is misaligned and fails to focus on the target. During horizontal gaze, the centroid motion in exotropia is more posterior compared to individuals with esotropia or controls.[13] Economides et al.[14] found that in strabismic individuals, the positional stability of both eyes is compromised compared to orthophoric individuals, indicating that strabismus impairs the ability to maintain steady fixation. Interestingly, our study did not observe significant differences in binocular DVA between the IXT and control groups at any viewing distance. This aligns with findings by Wang et al.,[15] who reported that decreased binocular DVA is more strongly associated with significant myopic refractive error rather than factors such as dominant eye function or accommodation. In our study, the proportion of binocular DVA abnormalities decreased in both groups. The reduced variability in binocular acuity may account for this outcome. Similarly, Jorge et al.[16] reported that binocular vision acuity is often superior to monocular vision acuity, further supporting our findings. Moreover, we observed that the proportion of monocular and binocular DVA abnormalities was significantly higher at 3 m compared to other viewing distances in both groups. As part of standard vision screening, young children are often tested at near distances rather than at 3 m, which is considered a greater challenge. The inclusion of participants aged 4 to 15 years in this study may have influenced the results, as younger children are generally more accustomed to using their vision at near distances than at greater viewing distances.[17]

Our findings offer a thorough assessment of the spatial abilities in children with IXT, including contour integration and stereopsis, at varying viewing distances compared to healthy controls. Children with IXT exhibited a significantly higher proportion of contour integration abnormalities at a viewing distance of 1 m compared to controls. Previous studies have reported that individuals with strabismus experience a loss of configural sensitivity in both eyes, regardless of whether visual acuity is reduced. Additionally, even those with good visual acuity in both eyes and poor stereopsis often exhibit deficits in configural sensitivity.[18] Neurophysiological studies in monkeys have identified the primary visual cortex (V1) as being intimately involved in contour integration.[19] Moreover, neuroimaging studies in humans have demonstrated that both striate and extrastriate areas contribute to this process.[20] Eye movements have long been recognized as integral to visual perception.[21] Normal smooth pursuit and saccadic eye movements are critical for the development of contour integration and object perception.[22] However, in children with IXT, limited or atypical eye movements may disrupt normal information flow to mid-eccentricity-biased higher-order visual cortices, adversely affecting their development. Consequently, this disruption induces visual perceptual deficits, particularly in categories associated with these regions. Furthermore, ocular stability in children with IXT is reduced compared to that of healthy controls, indicating that strabismus impairs the ability to maintain steady fixation on a target. Saccades contribute to variability in deviation angles due to the reduced conjugacy of eye movements in individuals with strabismus. Interestingly, faces and text, which occupy a small portion of the central visual field, require only minor eye movements for proper processing. In contrast, common objects that dominate the mid-peripheral visual field depend on longer-distance voluntary saccades to bring them into focus. This distinction may explain why the IXT group exhibited a significantly higher proportion of contour integration abnormalities at 1 m but not at other viewing distances. Additionally, gaze control mechanisms, such as the vestibulo-ocular reflex (VOR), are essential for stabilizing objects of interest on the high-acuity portion of the retina during body motion. In IXT, the instability of eye positions further compromises this ability, exacerbating fixation difficulties. In terms of stereopsis, no significant differences were observed between the IXT and control groups at any viewing distance in our study. This finding is consistent with previous reports suggesting no obvious correlation between the angle of deviation and the level of stereoacuity.[23,24] These results underscore the multifaceted impact of strabismus on visual perception, particularly in spatial tasks, while emphasizing the need for targeted interventions to address the specific challenges related to contour integration.

PEP is a novel concept introduced by Zhao et al. in 2014,[25] which describes binocular alignment under conditions more reflective of physiological function. Unlike traditional methods such as the Hirschberg test and cover test, PEP is assessed using a computer-controlled perceptual evaluation system under dichoptic vision conditions. This approach measures eye position when an object is perceived and visual information is integrated into the cortex, offering a more physiologically accurate and precise evaluation. Previous studies have shown that even individuals with normal apparent eye position may exhibit PEP abnormalities. For instance, when a smaller fixation target (approximately 1°) is used, subtler ocular misalignments in non-strabismic patients can be detected. In our study, we opted for a medium-sized fixation test object: a circular target measuring 0.8° ×0.8°, accompanied by a radiating cross spanning 0.66° ×0.66°. This choice was made to capture general patterns of ocular misalignment in both IXT patients and normal controls, allowing for a broader and more inclusive assessment. Our results revealed that IXT patients exhibited significantly greater deviations in dynamic perceptual ocular alignment compared to the control group, likely due to their diminished ability to maintain stable eye position through fusion force, despite impaired convergence.[26] Interestingly, we also observed small deviations in dynamic perceptual ocular alignment in the control group. Yang et al. have previously suggested that such abnormalities in individuals with normal apparent eye position may be associated with the development of severe anisometropia.[27] In summary, PEP provides a novel and precise framework for detecting subtle changes in ocular alignment enhances its clinical utility, offering valuable insights into the mechanisms underlying binocular vision deficits in IXT and other conditions.

Our study is the first to demonstrate that visual and perceptual examinations can effectively identify deviations in PEP values in both IXT and control groups, providing a sensitive and specific tool for assessing binocular functions that surpasses traditional testing methods. According to the ROC curve analysis for visual and perceptual examinations at various distances, different PEP thresholds exhibit compelling sensitivity and specificity for interpreting contour integration and stereopsis in children with IXT or in normal controls. In the control group, vertical PEP displayed high sensitivity and specificity for evaluating contour integration and stereopsis, aligning with previous findings that reported significant vertical PEP deviations in individuals with severe anisometropia without amblyopia or strabismus.[28] Moreover, prior research has demonstrated an association between fixation instability and deficits in stereoacuity, suggesting that fixation instability increases as stereoacuity decreases.[29,30] Similar deficits in contour integration have been observed in children treated for anisometropic amblyopia.[31] These findings may be attributed to disruptions in neural connectivity and the suppression of temporalward neurons within major visual areas, including V1, V2 (prestriate cortex), the medial temporal (MT), and medial superior temporal (MST) areas, extending to the posterior parietal cortex (PPC), a region critical for selective spatial attention.[32] In our study, both horizontal and vertical PEP demonstrated high sensitivity and specificity in interpreting contour integration in the IXT group. Individuals with good visual acuity in both eyes but poor stereopsis (nonamblyopic strabismus) also display configural sensitivity impairments, underscoring the pervasive impact of strabismus on binocular vision.[33] Our study underscores the utility of PEP as a robust marker for assessing visual and binocular functions and provides valuable insights into the neural mechanisms underlying visual deficits in IXT and related conditions.

In the present study, we utilized VR technology to evaluate binocular functions at varying viewing distances in children from both the IXT and control groups. The results revealed that both groups exhibited poorer visual functions at a far distance (3 m) compared to near (0.4 m) and intermediate (0.7 m and 1 m) distances. These findings, consistent with previous research, highlight the impact of poor stereopsis caused by childhood strabismus disorders on everyday perception and actions, particularly within near and intermediate visual spaces.[34] The diminished ability to perceive depth accurately in these contexts may be attributed to a lack of binocular depth information. This deficiency affects the midair localization of targets, making accurate spatial judgments challenging, even in individuals who are neither strabismic nor monocular. Such results underscore the importance of addressing binocular vision impairments to improve functional outcomes in daily life activities.

Conclusion

This study evaluated the effects of VR-based visual function assessment in children with IXT and healthy controls across different viewing distances. To our knowledge, it is the first to analyze the relationship between PEP deviation and binocular vision function in children with IXT and normal controls. While this research provides valuable insights, a deeper understanding of how sensory and tactile operations influence neural processes is essential. Further studies are warranted to explore the impact of VR-based interventions on brain activation patterns and to examine how various training parameters influence long-term changes in neural function. Such investigations will be critical in guiding future clinical applications and optimizing therapeutic strategies for binocular vision impairments.

Conflicts of interest:

There are no conflicts of interest.

Funding Statement

This research was funded by grants from the Science and Technology Program of Guangzhou, China (2023A04J1891).