Advances in behavioral vision training for the treatment of childhood amblyopia: a narrative review

Introduction

Amblyopia is a common neurodevelopmental visual disorder that arises in childhood. It is characterized by reduced best-corrected visual acuity (BCVA) in one or both eyes in the absence of detectable organic eye disease. This deficit results from abnormal early visual experience and is often disproportionate to the magnitude of the inciting visual anomaly (1). Amblyopia is a leading cause of monocular visual impairment in children, with an estimated global prevalence of 1.27% to 1.46% (1), and the number of affected individuals is projected to reach 175.2 million by 2030 (2). The prevalence of amblyopia varies significantly by geography. A meta-analysis reported higher rates in Europe and North America, at approximately 2.66% and 1.95%, respectively. In contrast, lower rates were observed in Asia (approximately 1.16%) and Africa (approximately 0.38%) (1). A cross-sectional study of urban and rural preschool children aged 30–83 months in central South China found an overall prevalence of 1.09% (3), consistent with the regional estimate for Asia. The core pathology of amblyopia is not a structural ocular defect. Rather, it is a functional disturbance of the primary visual cortex (also known as V1 area) and its downstream pathways induced by abnormal sensory input during the critical period of visual development (4). Without timely intervention, these disturbances can produce long-lasting deficits, including permanent reductions in visual acuity and impaired stereopsis (5). The consequences extend beyond acuity loss to involve higher-order visual processing and everyday functional skills. Amblyopia may be underrecognized and can impose lasting limitations on psychosocial well-being. The expert consensus therefore stresses early screening, prompt diagnosis, and timely intervention, as this can better exploit the sensitive period in visual development (6).

In clinical practice, standard amblyopia treatment protocols are primarily designed to promote visual function recovery in the affected eye. Common approaches include optical correction to eliminate defocus, occlusion therapy to enforce the use of the amblyopic eye, and optical or pharmacological penalization (e.g., atropine therapy) (7). Current guidelines recommend initiating treatment with full refractive correction to address any refractive component. An optical adaptation period is often advised, as this alone can lead to visual acuity improvement in some children. Occlusion therapy, which suppresses input to the fellow eye to force use of the amblyopic eye, remains the most longstanding and widely adopted strategy. A meta-analysis of 23 randomized controlled trials (RCTs) demonstrated that traditional therapies, such as part-time patching (2–6 hours daily) and atropine penalization, yield broadly comparable overall effects in improving amblyopic eye BCVA (5). However, these conventional treatment models have notable limitations. Both occlusion and penalization therapies are heavily dependent on patient adherence and the consistency of parental supervision. This reliance frequently results in poor compliance, prolonged treatment durations, and high recurrence rates. Extended patching can also cause ocular discomfort, significant psychosocial distress associated with wearing a patch, and growing resistance from the child. Consequently, many children struggle to complete interventions of sufficient intensity. In a Cochrane systematic review, definitive evidence indicating that any behavioral or technological intervention reliably improves objective adherence to occlusion therapy in children was lacking (8). Furthermore, traditional therapies primarily target monocular visual acuity, but they are less capable of restoring more complex visual functions, such as binocular integration and stereopsis. This limitation underscores the complex nature of amblyopia as a disorder involving binocular coordination and cortical development within the V1 area.

In recent years, the understanding of the neural mechanisms underlying amblyopia has advanced substantially. Under the traditional framework, visual plasticity is largely confined to an early critical period and declines rapidly with age (4). However, more recent evidence suggests that the V1 can undergo functional reorganization even beyond this window. Such reorganization can be induced by active, task-oriented visual learning that requires sustained attention. This novel understanding provides a neurobiological rationale for treating amblyopia across a broader age range (9). Against this background, behavioral vision training has emerged as a promising complementary or alternative therapeutic strategy. Grounded in principles of neural plasticity, this approach engages patients in repeated, carefully designed visual tasks, such as spatial localization, to strengthen specific visual functions (10). The central aim is to recruit the patient’s active participation to stimulate the pathway of the amblyopic eye and to promote binocular coordination. Two main technological trajectories have developed: perceptual learning and video game-based training. Perceptual learning, based on a method developed in a laboratory setting, involves high-repetition exercises that target basic visual features, such as orientation discrimination, to elicit long-lasting, task-specific improvements (11). Meanwhile, video gamified training offers greater ecological validity and patient engagement. This strategy is exemplified by dichoptic training, which entails the use of game platforms and specialized displays to present complementary stimuli to the two eyes and thereby promote binocular cooperation (12). Current clinical practice guidelines from the American Academy of Ophthalmology and the American Association for Pediatric Ophthalmology and Strabismus (AAPOS) recommend occlusion or atropine as first-line therapies for moderate amblyopia, with binocular digital therapies considered investigational except for specific devices that have received regulatory approval (7). Recent technological advances—including the widespread availability of consumer-grade virtual reality (VR) headsets, tablet-based gamified platforms, and cloud-connected home training systems—have substantially lowered the barriers to implementing behavioral vision training outside specialized clinical settings. These developments, together with growing recognition that conventional patching often fails to restore binocular function, make this an opportune moment to critically appraise the evidence for behavioral approaches.

A systematic synthesis of the literature on behavioral vision training for pediatric amblyopia, including its theoretical foundations, technical modalities, and available clinical evidence, can inform both clinical practice and further scientific investigation in this evolving field. A review was conducted that consolidated the recent research on behavioral vision training, evaluated its potential therapeutic role in amblyopia management, and examined the unresolved questions and future directions in this field. We present this article in accordance with the Narrative Review reporting checklist (available at https://tp.amegroups.com/article/view/10.21037/tp-2026-0464/rc).

Methods

A literature search of the PubMed database was conducted in December 2025 and was supplemented by a search of the Web of Science database to ensure comprehensive coverage. The search scope was limited to literature published between January 1, 2015 and December 31, 2025 (Table 1) to ensure collection of the most relevant and up-to-date evidence on visual training for children with amblyopia. This search was limited to literature published after January 1, 2015, primarily due to the commercialization of consumer VR headsets (such as the Oculus Rift CV1 released in 2016) and the widespread adoption of tablet-based binocular vision games. Earlier literature mainly focused on perceptual learning in adults, and related topics have already been reviewed in other literature.

A narrative review approach was chosen over a systematic review or meta-analysis for the following reasons. First, the interventions under consideration—perceptual learning, dichoptic training, gamified digital therapeutics, and VR-based training—differ substantially in task content, session duration, frequency, total treatment dose, and control conditions, making direct quantitative comparison problematic. Second, outcome measures vary across studies, including different visual acuity charts [logarithm of the minimum angle of resolution (logMAR) vs. Snellen], stereoacuity tests (Titmus, Randot, Frisby, TNO), and definitions of treatment success. Third, the included studies encompass a mix of RCTs, quasi-experimental designs, and single-arm case series, with varying risk of bias. Findings were organized by training modality and by outcome domain, with study design (e.g., RCT, prospective cohort, single-arm study) explicitly noted to help readers assess evidence strength. Divergent findings across studies were analyzed for potential sources of heterogeneity, including differences in patient age, amblyopia subtype, treatment dose, and control group intensity.

The search strategy combined controlled vocabulary (Medical Subject Headings terms) and free-text keywords related to the following domains: (I) population (keywords: “child”, “children”, “pediatric”, “paediatric”, “preschool”, and “school-age”); (II) condition (keywords: “amblyopia”, “lazy eye”, “anisometropic amblyopia”, “strabismic amblyopia”, and “refractive amblyopia”); (III) intervention (keywords: “behavioral vision training”, “perceptual learning”, “dichoptic training”, “binocular therapy”, “vision therapy”, “visual training”, “video game”, “gamified”, “digital therapeutic”, “virtual reality”, and “augmented reality”); and (IV) outcomes (keywords: “visual acuity”, “stereoacuity”, “stereopsis”, “contrast sensitivity”, “binocular function”, “adherence”, and “compliance”). The Boolean operators “AND” and “OR” were used to combine search terms appropriately. A manual search of the reference lists of included studies and relevant systematic reviews was also performed to identify additional eligible publications.

The inclusion criteria for studies were as follows: (I) RCTs, prospective or retrospective cohort studies, case-control studies, and single-arm interventional studies that specifically evaluated behavioral vision training interventions in pediatric amblyopia; (II) inclusion of children aged 0 to 18 years and diagnosed with unilateral or bilateral amblyopia of any etiology (anisometropic, strabismic, mixed, or deprivation) regardless of sex; (III) examinations of interventions involving any form of behavioral vision training, including perceptual learning, dichoptic therapy, gamified digital training, VR-based training, or combination approaches with occlusion; (IV) reported outcomes including at least one of BCVA, stereoacuity, contrast sensitivity (CS), binocular function parameters, adherence rates, or adverse events; and (V) published in English with full text availability.

Meanwhile, the exclusion criteria were as follows: (I) nonoriginal research, including review articles, conference abstracts, editorials, expert opinions, and case reports with fewer than five participants; (II) unclear research designs or insufficient data for extraction; (III) an exclusive focus on adult populations without disaggregated pediatric data; (IV) a lack relevant outcome measures related to visual function or treatment adherence; and (V) duplicate publications of the same cohort without additional novel data.

Data extraction was performed independently by two reviewers (Y.D., X.Y.) using a standardized form that included first author, publication year, study design, sample size, participant age and amblyopia type, intervention details (modality, duration, frequency, total dose, supervision setting), outcome measures (BCVA change, stereoacuity, CS, adherence, adverse events), and follow-up duration. Disagreements were resolved by discussion or adjudication by M.J.

Theoretical basis and principal forms of behavioral visual training

Etiology of amblyopia and basis of neuroplasticity

The etiology of amblyopia is multifactorial, and its classification is based on the type of abnormal visual experience that occurs during the critical period of development. Clinically, amblyopia is grouped into four principal types according to the underlying causative factor. Refractive amblyopia is the most common form and can be subdivided into anisometropic and iso-ametropic (bilateral high refractive) subtypes. Anisometropic amblyopia results from a substantial difference in spherical or cylindrical refractive power between the two eyes, commonly defined as a spherical difference ≥1.50 D or a cylindrical difference ≥1.00 D. This interocular disparity produces persistent inequality in retinal image clarity and size (13). The unequal input drives abnormal binocular competition for limited cortical resources, with the visual system selectively suppressing input from the more blurred eye and thereby impeding its visual development (13,14). Iso-ametropic amblyopia arises when both eyes have uncorrected, high refractive errors, such as marked hyperopia, myopia, or astigmatism. Under this condition, chronically blurred retinal images occur bilaterally, leading to a generalized delay in visual system maturation.

The pathophysiological mechanism of strabismic amblyopia is directly related to misalignment of the visual axes. This misalignment causes the image of a single object to fall on noncorresponding retinal points in the two eyes, for example, on the fovea in the fixing eye and outside the fovea in the deviating eye. Such discordant retinal stimulation produces diplopia and visual confusion (15). To eliminate this perceptual conflict, the V1 actively and selectively suppresses the visual signals originating from the deviating eye. This persistent monocular suppression constitutes the primary neural mechanism underlying the development of amblyopia. Amblyopia resulting from congenital or early-acquired obstruction of the visual pathway is termed “form-deprivation amblyopia”. Common causes include congenital cataract, ptosis, corneal opacity, and other similar conditions. This form of amblyopia is typically the most severe, as it severely or completely deprives the retina of patterned visual stimulation during development (16). This often leads to profound visual acuity loss and a range of complex visual functional deficits (17). Additionally, the related treatment involves iatrogenic risk, as inappropriate or excessive long-term occlusion or pharmacological penalization of the sound eye can potentially induce occlusion amblyopia or reverse amblyopia (18,19). This phenomenon underscores the vulnerability and plasticity of the developing visual system and highlights the critical need for close monitoring during traditional occlusion therapy.

The development of the visual system is governed by a temporally constrained window of heightened experience-dependent sensitivity and is considered to be a critical or sensitive period. This interval entails a finite timeframe during which the structure and function of the developing visual nervous system show pronounced plasticity in response to environmental input (20). Normal sensory experience is required for proper formation of cortical circuits. By contrast, abnormal visual experiences such as input deprivation, imbalanced binocular competition, or severely degraded image quality can rapidly and severely disrupt the maturation of the V1 (21,22). This may produce maladaptive changes, including abnormal encroachment of inputs from other sensory modalities, weakened functional connectivity between cortical layers, and reduced interhemispheric communication (23). In humans, the critical period for binocular vision and development of ocular dominance (OD) columns is thought to span the first few months of life up to 8 to 10 years, with the greatest plasticity concentrated in the first 2 to 3 years (6,20). Contemporary neuroscience has refined this view and confirmed that the V1 and higher-order visual areas retain residual experience-dependent plasticity into adulthood, albeit at a lower magnitude and slower pace than during early development (24). This residual plasticity underpins the neurobiological rationale for the application of amblyopia treatments beyond childhood, including behavioral vision training.

Neural mechanism studies have demonstrated that training with specifically designed visual tasks can modulate the response properties of neurons from the V1 to higher-order visual areas (25). This process also helps to balance the excitatory and inhibitory neurotransmission between the two eyes, thereby reducing the cortical suppression of the amblyopic eye and strengthening its neural representation. Research by Qin et al. indicated that in adult mice, the dorsal lateral geniculate nucleus (dLGN), the thalamic relay station for visual information, possesses intrinsic plasticity (26). Furthermore, this dLGN plasticity is essential for OD plasticity induced by monocular deprivation in the adult V1. In adult humans, brief periods of monocular patching can trigger rapid topographic reorganization within the cortical representation of the blind spot in V1. This region begins to respond to stimuli presented outside the blind spot, challenging the traditional notion of a fixed and unchanging V1 in adulthood (27). At the behavioral learning level, when adult macaque monkeys learned a contour detection task, structural remodeling was observed in the axon terminals of long-range horizontal connections within V1. These changes, which occurred specifically in neurons corresponding to the trained visual field location, included the sprouting of new axons and the pruning of old ones. This finding directly links behavioral learning to structural plasticity within cortical microcircuits (28). Similarly, in the V1 of patients with early-stage glaucoma, shifts and expansions in the receptive fields of local neuronal populations have been observed. This suggests the presence of localized, adaptive neural changes even in the context of adult-onset visual pathology (29).

Inhibitory interneurons, particularly those expressing parvalbumin, are critical regulators of plasticity. For instance, research has shown that inhibiting cyclin-dependent kinase 5 (Cdk5) can reinstate OD plasticity in the adult V1 (30). Furthermore, blocking oligodendrocyte differentiation during adolescence enhances the plasticity induced by monocular deprivation in adulthood (31). Acquafredda et al. discovered that the strength of functional connectivity between the pulvinar nucleus of the thalamus and V1 is negatively correlated with the degree of OD shift induced by short-term monocular deprivation in adult humans. This suggests that the pulvinar may act as a gatekeeper for V1 plasticity (32). An active visual training process requires sustained attention, decision-making, and anticipatory engagement from the patient. This engagement involves the top-down modulation of primary visual processing areas by higher-order cognitive cortices, such as the prefrontal cortex (33). This top-down modulation is thought to enhance the synchrony of activity within relevant neuronal populations and facilitate the release of neuromodulators (e.g., acetylcholine, norepinephrine). These changes, in turn, create a favorable chemical environment for synaptic plasticity. Therefore, behavioral vision training can be understood as a form of targeted neurorehabilitation. Through systematic, task-oriented, and attentionally demanding repetitive visual exercises, the response properties and connection efficiency of V1 neurons can be specifically and enduringly altered.

Principal types of behavioral visual training

Monocular perceptual learning-based training

Monocular perceptual learning is a behavioral vision training approach grounded in the principles of neural plasticity and has garnered substantial attention within the research community. This method requires patients to repeatedly perform specific visual tasks with their amblyopic eye, with task difficulty calibrated near their current perceptual threshold. The objective is to stimulate and reshape the brain’s visual processing pathways, thereby improving the function of the affected eye (34). This approach is particularly suitable for children who respond poorly to traditional occlusion therapy or who experience adherence issues. When the amblyopic eye is consistently challenged to discriminate visual stimuli at the limits of its capability, the specific neural pathways engaged in processing that task are selectively reinforced. This typically manifests as a systematic decrease in perceptual thresholds, such as the minimum detectable contrast (35). This improvement does not stem from any change in the eye’s optical structure. Rather, it reflects an enhanced capacity of the brain’s visual centers to filter and interpret blurred or noisy signals. This indicates increased neural connection efficiency along the pathway from the LGN to the V1 (36).

In practice, the design of training paradigms typically targets the specific neural deficits commonly associated with amblyopia. Core tasks include CS detection, grating orientation discrimination, and visual motion direction discrimination (34). One study found that after patients with amblyopia underwent training on contrast detection of central targets (e.g., Gabor patches), their improvements were not confined to CS alone. These gains broadly transferred to other spatial tasks, including visual acuity, vernier acuity, and foveal crowding tasks (37). In another study, monocular CS perceptual learning significantly enhanced both CS and the balance point in the trained eye of anisometropic amblyopes at the trained spatial frequency (six cycles per degree). These effects also generalized to adjacent, untrained spatial frequencies, although no significant interocular improvement was observed (35). Orientation discrimination and spatial localization tasks primarily target the deficits in contour integration and shape perception common in the amblyopic eye. For example, patients may be asked to judge the tilt orientation of a grating or to locate a target line segment among a field of distracting lines (38,39). This type of training aims to enhance the visual system’s ability to encode object boundaries and its resilience to visual noise.

Notably, the effects of perceptual learning demonstrate a degree of task generalization. For example, training focused on spatial CS not only improves static visual acuity but may also confer benefits to dynamic visual functions. A study involving children who had undergone surgery for esotropia found that home-based perceptual learning significantly enhanced CS at medium and high spatial frequencies and facilitated the recovery of stereopsis (40). Another study reported that training based on a lateral masking paradigm improved stereopsis in individuals with amblyopia (41). Training paradigms have continued to evolve in recent years. For instance, perceptual learning utilizing temporally modulated flicker stimuli has been shown to effectively improve the critical flicker fusion frequency in adults with amblyopia. Furthermore, these gains are transferable to other spatial and binocular functions, including visual acuity, CS, and flicker-defined stereopsis (42).

In one study, patients with anisometropic amblyopia completed a monocular fine orientation discrimination perceptual learning task. During training, both the spatial frequency and orientation of the stimuli were individually calibrated to each participant’s perceptual threshold (43). This training was found to both improve patients’ BCVA and provide marked enhancement in their uncorrected visual acuity (E-chart vision with an average improvement of approximately 1.3 lines). Concurrent improvements in CS and stereopsis were also observed (43). This constitutes evidence that targeted training for higher-order visual processing deficits can effectively generalize to improvements in basic visual acuity. In clinical practice, the design of monocular perceptual learning paradigms is typically characterized by adaptively adjustable task difficulty and explicit feedback. By consistently requiring the participant to perform tasks near their individual threshold, these designs sustain the learning drive. This feature is critically important for enhancing long-term patient adherence and ensuring the effectiveness of the therapeutic intervention.

Binocular vision rehabilitation-based training

Traditional occlusion therapy primarily aims to improve monocular visual acuity in the amblyopic eye but often fails to effectively restore binocular visual function (9). Training based on binocular vision reconstruction represents a paradigm shift in amblyopia treatment, with active, interactive models that promote binocular cooperation becoming favored over conventional monocular occlusion (10). The core objective of this approach is to leverage meticulously designed visual tasks to compel or encourage simultaneous and coordinated use of both eyes, thereby rebuilding normal binocular visual function, including fusion, accommodation, and stereopsis. The theoretical foundation of this training model stems from a growing understanding of the neurophysiology of amblyopia. It is now recognized that amblyopia involves not merely a reduction in monocular visual acuity but—more fundamentally—a disruption or inhibition of the binocular visual pathways, particularly the higher-order cortical functions involved in fusion and stereopsis (44).

Dichoptic training is the most actively investigated area within this field. This method uses dichoptic presentation techniques, such as red-blue or polarized glasses, to display images with differences in contrast, luminance, or content to each eye. Reducing the salience of the input to the fellow eye, for instance, through applying a Gaussian blur, or assigning complementary elements within a game-based tasks can compel the two eyes to work together to accomplish an overall visual task. This process is intended to gradually reduce interocular suppression during gameplay or interactive engagement. For example, one study employed a polarizing film dichoptic method and instructed children with anisometropic amblyopia to perform game tasks that necessitated binocular integration. The improvement in visual acuity in the polarizing film group was 2.1 times greater than that in the eye patch group (45). More recent research has examined dynamic adjustment of dichoptic stimulus parameters, such as modulating contrast based on the patient’s real-time performance, for optimizing the training process. One study used eye-tracking technology to monitor the eye movement patterns of patients with amblyopia and healthy controls while they viewed dichoptic videos and stationary dots with varying levels of fellow-eye contrast (46). The study found that reducing the contrast of the image presented to the fellow eye effectively increased the duration of fixation by the amblyopic eye within the amblyopic zone. This effect was particularly pronounced in patients with smaller angles of strabismus and in those who had undergone surgical correction for strabismus. Conversely, in patients with larger angles of strabismus, the degree of ocular deviation increased as the contrast in the fellow eye was reduced.

Another approach involves binocular balance and suppression inhibitory training. Methods with this design focus more on recalibrating the neural excitation balance between the two eyes within relatively static visual tasks. Instruments such as the synoptophore, vergence balls, and red-green vectograms are employed for targeted exercises in simultaneous perception, fusional vergence ranges (convergence and divergence), and stereopsis. The goal is to enhance binocular coordination and stability during dynamic visual processes (47). Research indicates that even in children with a history of occlusion therapy, such dichoptic perceptual training can effectively promote the recovery of stereopsis. This effect is particularly pronounced in children with mild amblyopia, defined as an amblyopic eye visual acuity of ≤0.28 logMAR (48).

Furthermore, stereopsis and fusion-specific training directly targets and seeks to ameliorate higher-order binocular visual functions. Training tasks use tools such as random-dot stereograms and vectograms for exercises in fusion range development, fine stereopsis, and the coordination of accommodation and vergence. For example, one study employed customized, computerized perceptual learning tasks for binocular vision training. Through repetitive application of adaptively challenging stimuli, this approach aims to drive plastic changes within relevant neural pathways (49). Such training is predicated on the principle of task-specific plasticity. The repeated practice of tasks that demand precise binocular coordination can directly reinforce the efficacy of the corresponding neural circuits. As the related technology advances, these traditional training principles are being deeply integrated with digital therapeutics. This has led to the development of interactive, gamified training programs delivered on platforms such as tablets or VR headsets. An 8-week study using a video game-based binocular training paradigm reported posttreatment improvements in stereopsis for 14% of participants and a reduction in interocular suppression for 43%. Notably, individuals with strabismic amblyopia exhibited more pronounced improvements in global motion perception (50). Other evidence suggests that even in patients who have responded poorly to traditional occlusion therapy or who are beyond the typical sensitive period for visual development, this active training model directly targeting interocular neural competition can still yield substantial improvements in both visual acuity and stereopsis (51).

Digital and gamified training modes

The digitalization and gamification of behavioral vision training delivery systems leverage principles of game design and digital interaction technologies to restructure the logic of treatment implementation and the user experience. The core objective of this approach is to enhance therapeutic motivation, engagement, and long-term adherence in pediatric patients. Although traditional occlusion therapy has demonstrated efficacy, it is often characterized by a lack of engaging content and monotonous treatment procedures. This frequently results in poor adherence among children, which can ultimately compromise the final therapeutic outcome. Computer-assisted vision training relies on personal computers or tablet devices. By designing visual tasks with adjustable parameters, this modality enables both personalization and standardization of training content (52). Gamification elements can be systematically integrated with this foundation. Specific visual tasks are embedded within a game narrative framework that features clear objectives, immediate feedback, reward mechanisms, and progressively challenging levels. This design transforms the therapeutic process from a tedious regimen of repetitive exercises into an engaging experience of exploration and challenge.

Research has demonstrated that dichoptic games developed for the iPad platform have the potential to improve stereopsis and visual acuity in children with amblyopia, with an efficacy comparable to that of traditional part-time patching (53). An RCT further confirmed that, in children with amblyopia, the therapeutic effect of gamified binocular treatment is similar to that of conventional patching and may offer advantages in certain visual function parameters (54). Studies have also found that various game formats, including falling blocks games, dig rush games, and roguelike shooting games, all contribute to mean visual acuity improvement in children with amblyopia (51). Among these, the roguelike shooting game employs red-blue dichoptic technology to present differentiated game elements, such as characters and monsters, to each eye. This design requires the child to coordinate both eyes to perform aiming and shooting tasks. This approach ingeniously transforms binocular fusion and stereopsis training into the core operational mechanism of the game, thereby providing a unique therapeutic advantage.

The introduction of VR and augmented reality (AR) technologies has enabled the creation of highly controllable immersive or mixed visual environments for behavioral vision training. VR technology, involving head-mounted displays, generates a completely immersive and enclosed visual environment. This allows for the precise and independent control of visual input to each eye and provides an ideal platform for dichoptic training. Preliminary research suggests that VR-based visual training programs hold potential value for inducing neuroplasticity in children with amblyopia (55). Furthermore, VR rehabilitation protocols for these patients have already progressed to the implementation phase in multicenter randomized studies (56), and a few VR-based therapies have received regulatory approval. In one such approach, software in a specific device converts content into therapeutic visual stimuli in real time and transmits them via a wireless network. The stimuli are then presented to the patient’s head-mounted display in a binocular parallax manner to reduce the contrast of the image received by the dominant eye and encourage binocular coordination (57). In contrast to the complete immersion of VR, AR technology integrates training seamlessly into real-world daily activities. Gao et al. developed a system that uses lightweight AR glasses to process and modify, in real-time, the video stream of the wearer’s actual surroundings (58). This technology can, for example, specifically disrupt the phase of low spatial frequency information received by the dominant eye while preserving high spatial frequency details. This allows for the targeted intervention of specific visual pathways, such as the magnocellular pathway, during natural behaviors such as walking or observing the environment. This approach opens new avenues for delivering highly efficient and ecologically valid amblyopia treatment within the context of everyday life.

Furthermore, home-based training platforms on smartphones or tablet computers constitute another critical dimension of digital therapeutics. These systems offer convenient, in-home training guidance for patients and their families, deliver personalized treatment regimens, and enable automatic recording and remote uploading of training data (59,60). For example, several studies have used cloud-based online platforms through which pediatric patients can perform multiple training sessions every week, with each lasting approximately 30 minutes. The findings indicate that this approach can improve both visual acuity and stereopsis (60). By incorporating gamified training modules, such platforms enhance treatment accessibility, particularly for patients in regions with limited healthcare resources. Meanwhile, the objective data tracking they provide establishes a reliable foundation for assessing both patient adherence and therapeutic efficacy. One exploratory study reported that a gamified mobile application could effectively enhance treatment adherence in children with amblyopia (61). In an RCT conducted by Uttamapinan et al., it was confirmed that using a specific smartphone application for adjunctive management could significantly improve BCVA levels in children with amblyopia while simultaneously increasing adherence to occlusion therapy (62). This model of remote monitoring and management enables clinicians to make timely adjustments and provide professional supervision of home-based training via the upload of objective data. Consequently, it offers the potential to establish a more efficient, patient-centered, closed-loop therapeutic management system. However, a meta-analysis has also noted that while binocular game therapy yields considerable visual acuity benefits compared to traditional patching in both the short and long term, a greater body of high-quality research evidence is needed to confirm the stability of these long-term effects (63). In summary, digital and gamified training modes, empowered by technology, are progressively transforming amblyopia treatment. They are shifting the paradigm from a passive, isolated medical intervention toward an active, integrated, and more engaging process of visual function remodeling.

Clinical efficacy of behavioral vision therapy for amblyopia in children

Multidimensional visual function outcomes

Historically, the assessment of clinical efficacy for amblyopia treatments, including traditional occlusion therapy, has focused primarily on improvements in monocular visual acuity. Such assessment now encompasses a multidimensional and systematic approach that can address a range of neurovisual deficits associated with amblyopia. Recent evidence indicates that a number of training protocols can lead to varying degrees of improvement in BCVA, binocular visual function, and higher-order visual quality.

Improvement in BCVA

BCVA is a key quantitative metric for assessing the degree of visual function recovery in children with amblyopia. It has long been used to evaluate treatment responses to different interventions and is frequently employed as a primary or coprimary outcome measure in studies on behavioral vision therapy. Table 2 summarizes the effects of behavioral vision therapy on BCVA. Evidence suggests that the therapeutic effect on the amblyopic eye varies depending on the training paradigm, technical approach, and study design. Monocular perceptual learning-based methods, for instance, have shown potential in improving BCVA. In one study, perceptual learning improved BCVA in the amblyopic eye from approximately 0.45 logMAR at baseline to 0.25 logMAR (64). After a short period of intensive training (e.g., 10 days), a statistically significant improvement in BCVA from baseline was observed (65). However, when compared with traditional occlusion therapy, no significant difference in BCVA change was found at the 1-month follow-up, suggesting some overlap in the visual acuity improvement trajectories between the two interventions (64). In certain amblyopia subtypes, such as meridional amblyopia, a monocular fine orientation discrimination perceptual learning (fine-PL) protocol has been applied, leading to improvements in both uncorrected visual acuity and BCVA (43). Overall, monocular training paradigms appear biologically plausible for improving BCVA. However, their relative advantage over other treatments has not been consistently demonstrated across different study designs and follow-up durations.

With a deeper understanding of the mechanisms underlying interocular suppression, dichoptic training and gamified digital therapy have emerged as key research foci. These approaches aim to improve amblyopia-related visual function by reducing the weight of signals from the dominant eye and re-establishing binocular integration. Several single-center or quasi-experimental studies have used cloud platforms, tablet computers, or specialized systems for home-based training. Participants in these studies were often children with anisometropic amblyopia or amblyopia associated with microstrabismus (49,51,60). Improvements in BCVA were frequently reported over the course of the training period. However, due to the frequent absence of a parallel control group in these study designs, the observed changes should be more appropriately interpreted as describing intervention-related characteristics rather than as rigorous evidence of comparative efficacy. In combined treatment contexts, incorporating perceptual learning or Cambridge Visual Stimulator (CAM) therapy as a supplement to occlusion has been observed to shorten the time required to achieve predefined treatment success criteria in children with moderate amblyopia (66). In contrast, one RCT found no significant difference in the mean BCVA between CAM therapy alone and traditional occlusion at any follow-up time point (67). Some studies suggest that specific technological approaches may offer relative advantages under certain conditions. For instance, in a comparison between VR-assisted training and conventional optical correction, the VR group achieved a higher rate of treatment success (68). Similarly, binocular therapy using polarizing films was reported to yield greater visual acuity improvements after 2 months of treatment as compared to traditional patching (45). Of note, the interpretation of BCVA improvements from single-arm studies requires caution. These designs lack a parallel control group and cannot account for spontaneous improvement or regression to the mean. In contrast, the larger RCTs summarized in Table 2 provide higher-certainty evidence, albeit with inconsistent findings across different age groups and training platforms.

It is worth noting that behavioral training may be superior to, or at least noninferior to, traditional occlusion therapy under certain conditions (76,77). For instance, a pivotal trial of an eye tracking-based home treatment system (CureSight) demonstrated noninferiority compared to traditional patching (69). Meanwhile, another RCT reported that in a per-protocol analysis, the improvement in BCVA was significantly greater with this system than with patching (70). In a proof-of-concept trial involving young children, a type of binocular gamified digital therapy (Vision Planet) was associated with continuous visual acuity improvements over the treatment period (51). However, in another RCT, this therapy only showed superior visual acuity gains compared to patching at the 1-month mark, with no statistically significant difference between the two treatment groups at 6 months (54). Similar inconsistencies have been observed in studies of other binocular games. An iPad game (Dig Rush) for children aged 4 to 6 years demonstrated superior visual acuity improvement compared to optical correction alone over the short term (4 weeks) (71). In contrast, larger studies involving children aged 7 to 12 years (72) and 5 to 13 years (53) found that such game-based training either provided no advantage or was less effective than traditional occlusion therapy. Furthermore, passive treatments that adjust the contrast of binocularly viewed content, such as dichoptic cartoons or movies, have also been shown to produce significant BCVA improvements within 2 weeks. Their effect was comparable to or, in some studies, better than that of patching (73,74). However, not all research supports the absolute superiority of behavioral training. In several comparative studies, the difference in BCVA improvement between the behavioral training and patching groups was not statistically significant at the treatment endpoint (53,54,67,78,79). This suggests that the two approaches may be broadly comparable in terms of the final visual acuity outcome.

VR-based training leverages immersive environments to enhance dichoptic stimulation. Its effects on BCVA, however, have been inconsistent across studies. Several investigations have reported significant improvements in amblyopic eye BCVA following gamified VR approaches, including those using the interactive binocular treatment (I-BiT) system, novel VR platforms, and the Luminopia One system (80-82). In a phase III RCT, the Luminopia One system demonstrated superior visual acuity gains compared to refractive correction alone (57). Some RCTs have reported statistically significant within-group changes in BCVA for both VR training and patching, but between-group comparisons did not reveal a consistent advantage for either intervention (75,83). In summary, behavioral vision therapy, as a class of interventions, can effectively improve BCVA in the amblyopic eye. However, its superiority relative to traditional standard treatments, such as occlusion, is likely highly dependent on the specific technological implementation, the characteristics of the study population, the treatment dose and duration, and the intensity of the control intervention.

Binocular function and other visual quality metrics

Compared to BCVA, which is based on the visual acuity chart, binocular function and related visual quality parameters better reflect the impact of amblyopia interventions on the overall integration and higher-order processing capabilities of the visual system. Consequently, they are being increasingly recognized as important outcome dimensions in studies on behavioral vision therapy. Table 3 summarizes the effects of behavioral vision therapy on selected binocular and visual function metrics. In previous studies, stereopsis has often been assessed with tools such as the Titmus, Randot, Frisby, VacMan, and TNO (Netherlands Organisation for Applied Scientific Research) tests (49,51,79,86). Within the field of dichoptic training, multiple studies have reported statistically significant improvements in stereoacuity following interventions. These include treatments based on online cloud platforms (e.g., Bynocs) (60), VR-based dichoptic games (75), and specifically designed gamified training programs (51,54,69,85). However, findings across studies have been inconsistent regarding direct comparisons to traditional occlusion therapy, and most between-group comparisons have failed to demonstrate a clear advantage for dichoptic training over patching (54,69,70,75). Several factors may explain these inconsistencies. First, stereoacuity measurement tools vary (Titmus, Randot, Frisby, TNO) and have different sensitivity and test-retest reliability. Second, baseline stereoacuity differs substantially across study populations; children with measurable baseline stereopsis are more likely to show improvement than those with nil stereopsis. Third, the duration of dichoptic exposure—ranging from 2 weeks to 6 months—may determine whether binocular gains reach statistical significance.

In addition, several studies have examined the effects of short-term monocular or binocular training on stereopsis. Findings indicate that perceptual learning can lead to significant improvements in stereopsis among children with amblyopia aged 4 to 12 years (36,65). Active vision therapy (AVT) has also been reported to yield greater gains in stereopsis compared to occlusion therapy (84). A prospective multicenter pilot study investigating a novel VR system observed significant posttreatment improvements in both stereoacuity and the break and recovery points of distance negative fusional vergence. This approach also demonstrated favorable applicability in older children with amblyopia (80). Multiple studies on combination therapies suggest that integrating behavioral training with interventions such as Fresnel prism correction or patching can lead to positive changes in binocular function scores and stereopsis during follow-up. However, the degree of improvement appears to be substantially influenced by baseline binocular status and treatment adherence (59,77,87,88). Moreover, improvements in stereopsis have not been universally observed. For instance, some passive video-based treatments that rely on contrast rebalancing have been shown to improve visual acuity without producing significant gains in stereopsis (73,74). This suggests that the pathways through which different training mechanisms influence binocular function may vary.

Behavioral vision training has demonstrated positive effects on other visual quality metrics. CS is a key parameter for assessing visual function. Studies have shown that perceptual learning-based training can improve CS in the amblyopic eye across multiple spatial frequencies (59,64). When compared to other intervention types, dichoptic video game therapy was found to be superior to traditional occlusion for improving CS (78). However, studies employing VR-based dichoptic training did not observe significant changes in CS (85). Beyond CS, such training can also improve other functions, such as accommodative facility. One study using a VR serious game reported a significant increase in mean accommodative facility in the amblyopic eye, from 1.44 cycles per minute to 4.96 cycles per minute (85). This suggests that immersive training may positively influence near visual accommodation. However, findings regarding the resolution of interocular suppression and the improvement of interocular balance remain inconsistent. One study found that binocular balance vision (BBV) therapy improved visual acuity but did not significantly alter the state of interocular suppression as measured by the VacMan test (79). In contrast, another study on visual perceptual learning reported improvements in perceptual eye position and a significant increase in the proportion of children without interocular suppression following training (65). These findings indicate that the scope of behavioral vision training extends beyond the recognition of optotypes on a visual acuity chart and may influence the underlying neural processes that shape everyday visual experience.

Comparison with traditional therapy and combination strategies

In the clinical management of childhood amblyopia, occlusion therapy has long served as the standard treatment. Its efficacy is closely related to age, adherence, and the type of amblyopia. With a deeper understanding of the binocular mechanisms underlying amblyopia, behavioral vision training is being more frequently introduced as a first-line alternative, a component of combination therapy, or a salvage strategy for cases with an insufficient response to traditional treatment. Table 4 provides examples of studies comparing behavioral vision training with traditional occlusion therapy and those investigating combination approaches. Behavioral training often demonstrates effects comparable to traditional occlusion therapy in improving BCVA in the amblyopic eye. For instance, evidence suggests that there are no significant differences in mean posttreatment BCVA between CAM or AVT and occlusion therapy (67,84). A few rigorously designed noninferiority trials have confirmed that certain interventions, such as home-based therapy with the CureSight system, are not only noninferior but, in some analyses, superior to traditional patching (69,70). Similarly, a binocular vision training program for young children showed superior visual acuity improvement compared to patching alone within a short, 1-month period (54). However, not all evidence supports this equivalence. One large, multicenter study involving children aged 5 to 13 years found that traditional occlusion therapy resulted in a significantly greater BCVA improvement than did binocular iPad game training after 16 weeks of intervention (53). This suggests that factors such as the design of the training protocol, patient characteristics, and treatment duration and intensity may be critical variables influencing the outcomes in studies comparing the efficacies of interventions.

Beyond direct comparisons of efficacy, behavioral training and traditional therapy exhibit potential complementarity in terms of adherence, mechanisms of action, and therapeutic goals. This has prompted the exploration of combination treatment strategies. These approaches typically endeavor to integrate the forced use of the amblyopic eye induced by occlusion with the active stimulation of visual neural pathways provided by behavioral training, with the aim of achieving synergistic benefits. Several studies have reported that combining monocular perceptual learning or child-friendly action video games with patching leads to greater visual acuity improvements than does patching alone (36,88). In other research, monocular video games or smartphone applications used in conjunction with patching resulted in superior visual acuity gains at both 1 and 3 months as compared to daily patching alone (62,89).

Combination therapy may also exert a positive influence on the restoration of binocular visual function. For instance, active vision training combined with occlusion has been observed to improve and maintain binocular visual function in affected children (87). The core rationale underlying these combination strategies is that occlusion first overcomes suppression of the dominant eye, after which behavioral training further promotes visual information processing in the amblyopic eye and the active reconstruction of binocular coordination. However, studies on combination therapy have not uniformly reported benefits from this approach. One quasi-experimental study found no significant difference in final visual acuity improvement between a group receiving perceptual learning and binocular therapy combined with occlusion and a group receiving traditional occlusion alone (59). These findings suggest that optimizing combination protocols, including factors such as the timing of intervention and the appropriate balance between training content and occlusion dose, represents a critical area for further investigation. In summary, the relationship between behavioral vision training and traditional occlusion therapy is not simply one of replacement. Behavioral training can serve both as a salvage option for children who are intolerant or poor responders to occlusion and as an adjunctive or sequential treatment aimed at promoting the recovery of binocular visual function. Together, these approaches constitute a therapeutic arsenal that can be flexibly combined and tailored to the individual characteristics of each child.

Proposed clinical decision pathway for behavioral vision training in childhood amblyopia (Figure 1). The algorithm begins with confirmation of amblyopia diagnosis and completion of full refractive correction (4–16 weeks). For children aged <4 years, passive dichoptic video or cartoon viewing (contrast-rebalanced) is preferred due to low cognitive demand. For children aged 4–7 years with moderate amblyopia (BCVA 0.3–0.7 logMAR) and no prior patching failure, gamified tablet-based dichoptic training (e.g., Dig Rush-type games) may be offered as a first-line alternative or in combination with part-time patching. For children aged ≥7 years with poor adherence to patching, persistent binocular dysfunction, or anisometropic/mixed amblyopia, home-based eye-tracking systems (e.g., CureSight) or supervised VR training may be considered. Combination with patching is reserved for cases with suboptimal response to monotherapy after 8–12 weeks. Regular monitoring (every 4–8 weeks) of visual acuity, stereoacuity, and ocular alignment is recommended. Treatment modification or discontinuation is advised if diplopia, worsening strabismus, or persistent headache occurs.

Figure 1 Proposed clinical decision pathway for selecting behavioral vision training modalities in childhood amblyopia. The algorithm begins with confirmation of amblyopia diagnosis and full refractive correction. For children aged <4 years, passive dichoptic video or cartoon viewing is preferred. For those aged 4–7 years with moderate amblyopia and good adherence potential, gamified tablet-based dichoptic training may be offered. For children aged ≥7 years with prior poor adherence to patching or persistent binocular dysfunction, home-based eye-tracking systems or supervised VR training can be considered. Combination with patching is reserved for cases with suboptimal response to monotherapy after 8–12 weeks. Regular monitoring of ocular alignment and subjective symptoms is recommended, with treatment modification if adverse events (diplopia, worsening strabismus) occur. VR, virtual reality.

Treatment adherence and adverse effects

Treatment adherence is a critical factor influencing the effectiveness and clinical applicability of amblyopia interventions in children. It is also a key dimension that is frequently discussed and compared in the evaluation of behavioral vision training relative to traditional occlusion therapy. Traditional patching is often limited by poor adherence in clinical practice, partly due to its impact on appearance, social activity, and comfort. This issue may be particularly pronounced in school-age children and those with mild-to-moderate amblyopia. Behavioral vision training attempts to reduce treatment burden and improve acceptance among children and caregivers by incorporating gamified interfaces, binocular interactive stimulation, and home-based application models. Methods for assessing adherence vary across studies. Common metrics include the actual duration of treatment completed, data automatically logged by the device, parental logs, and researcher follow-up assessments (53,71,72). Multiple studies have shown that digital and gamified training often achieves high completion rates in the short term. For example, in a clinically supervised setting, a VR serious game yielded mean completion and duration rates of 88.16% and 85.41%, respectively (85). Meanwhile, the home-based CureSight system yielded median adherence rates of approximately 94% in two independent studies, which were significantly higher than those in the respective patching groups. Caregiver satisfaction exceeded 90% in both studies (69,70). Other home-based training platforms, such as Bynocs and contrast-rebalanced cartoon videos, have also maintained high mean adherence rates ranging from 87% to 98% over intervention periods of 2 to 6 weeks (60,73).

Some studies suggest that embedding therapeutic tasks within game mechanisms or media content familiar to children can help enhance active participation (51,74). However, adherence data based on parental self-reports may be subject to recall bias or reporting bias (45,90). Furthermore, maintaining adherence over the long term presents inherent challenges. Some studies have observed a tendency for adherence to decline gradually as the treatment period extends (59,62). For instance, one study on combination therapy reported that adherence decreased from 72.4% in the first month to 56.8% in the third month (59). In addition, different recording methods, such as parental reports (71,72) and device logs (51,53), and different implementation settings, such as clinic-supervised training (91) and home-based training (57,60,64,82), may introduce systematic bias into adherence outcomes. In contrast, some studies on iPad-based binocular games and BBV have observed decreased treatment completion rates or participant dropout during follow-up despite the inherently engaging nature of the treatment protocols (53,72,79). These findings suggest that the potential adherence advantage of behavioral vision training is not absolute. Rather, it appears to be closely related to the design of the protocol, the implementation setting, the mode of interaction, and the presence of effective long-term incentive mechanisms, all of which should be further investigated.

In terms of safety, the potential adverse effects associated with behavioral vision training, particularly methods involving dichoptic stimulation and high accommodative demand, differ from those of traditional occlusion therapy. Based on available study reports, the most common adverse effects are mild to moderate in severity and typically transient. Headache and eye fatigue are relatively frequently reported symptoms. These have been noted with home-based CureSight therapy (70), iPad binocular games (71), and some forms of VR training (80). Risks directly related to the binocular visual state include diplopia, worsening of strabismus, or the new onset of strabismus. Such events have been reported in several studies involving iPad games (53,71,72), VR training (75), and BBV therapy (79). Beyond visual symptoms, studies on VR training have also reported systemic reactions associated with immersive devices, such as dizziness, physical fatigue, and sweating. However, these were generally mild to moderate in severity and did not typically lead to treatment discontinuation (75). Safety evaluations in controlled studies on digital therapeutics, such as CureSight and Luminopia One, which adhere more closely to the regulatory standards for medical devices, have systematically attested to the incidence of adverse events, but none were serious or irreversible (57,69,70). In contrast, some earlier studies on perceptual learning and combined patching did not systematically report adverse events, which limits a comprehensive assessment of their safety profile (36,64,65). One study specifically reported that no recurrence of visual acuity loss was observed during long-term follow-up after active vision training combined with occlusion (87).

The occurrence of these adverse effects may be related to the training stimulating the amblyopic eye—which has been under long-term suppression—to re-engage in binocular competition. It may also impose new adaptive demands on the accommodative and vergence systems. Although serious adverse events are rare, conducting a thorough pretreatment visual function assessment before the initiation of dichoptic training is considered a necessary clinical safety measure for children with latent strabismus or insufficient fusional reserves. Close monitoring of ocular alignment and subjective symptoms during treatment is also recommended. Table 5 summarizes the studies on the adherence and adverse effects associated with behavioral vision training for childhood amblyopia.

Other factors influencing treatment efficacy

The therapeutic response to behavioral vision training is influenced not only by the design of the training protocol itself but also by a combination of multiple factors, including individual patient characteristics and specific parameters of treatment implementation. Age is often regarded as a critical factor, as it is closely related to the temporal window of neuroplasticity in the visual system. Several studies have included participants across a wide age range, from 3 to 18 years (36,49,76,87). However, the bulk of research in this area has focused on preschool and early school-age children, such as those aged 3 to 5 years (73), 4 to 8 years (54), or 5 to 12 years (65). For instance, binocular digital therapies, VR interventions, or animated video training for children aged 4 to 7 years often adopt a home-based model with high frequency but short session durations. This approach aims to reduce fatigue and improve acceptability (57,71,73). Similarly, research on VR video game therapy found that children under 5.5 years of age could not understand the game setup or the game itself. In contrast, older children were less willing to comply with the gaming protocol. This suggests that loss of interest in the game may be a limiting factor for treatment across all pediatric age groups (91). It is worth noting that even within a relatively broad age group, age differences may contribute to heterogeneity in outcomes. In a study on combination therapy, the mean age of the visual therapy group was significantly higher than that of the traditional patching group, which might have confounded comparisons of efficacy between the groups (59). For older children, such as those aged 8 to 15 years, novel VR systems have also demonstrated potential for improving visual acuity and binocular function. However, these studies are mainly small, exploratory trials (80). Collectively, these findings suggest that while neuroplasticity diminishes with age, behavioral vision training may offer additional therapeutic opportunities for children who are beyond the optimal age window for traditional occlusion therapy.

The specific type of amblyopia and its severity also significantly influence the application scenarios and potential efficacy of behavioral vision training. Research has examined the main amblyopia types, including refractive, anisometropic, strabismic, and mixed amblyopia. Some studies have focused on specific etiological subgroups, such as children with anisometropic amblyopia with or without microstrabismus (60) or those with esotropia and amblyopia (77). However, the majority of studies have included populations with mixed etiologies to assess the broad applicability of the training methods (53,54,59,79,84). Different types of amblyopia may exhibit distinct sensitivities to various training approaches. For example, one study found that dichoptic video treatment based on contrast rebalancing resulted in superior short-term visual acuity improvements as compared to traditional patching in children aged 3 to 5 years with strabismic or anisometropic amblyopia (73). However, another similar study in children aged 3 to 7 years found no significant between-group difference in visual acuity improvement after 2 weeks of the same video therapy when compared to patching (74). It has also been suggested that patients with more severe initial deficits may derive greater benefit from certain treatments, such as progressive AVT (88). Conversely, in studies on BBV, it was found that baseline differences in stereoacuity across patient subgroups may complicate the interpretation of whether efficacy differences are due to treatment effects or sample selection bias (79). Regarding amblyopia severity, most of the related studies have enrolled children with moderate amblyopia, and others have explicitly excluded children with severe amblyopia or controlled for it by stratifying patching intensity and training duration (76,79,84). This selection, to some extent, limits the generalizability of findings from behavioral vision training to severe amblyopia or complex clinical situations. It also reflects the practical considerations of researchers regarding safety and adherence.

Training intensity and frequency are among the most significant sources of heterogeneity in studies on behavioral vision training. In the research on this subject, training frequency ranges from several times per week (67,68) to every day (60,69,71), session duration ranges from 20 minutes (76,77) to 2 hours (78,84), and total intervention period ranges from 2 weeks (73,74) to 6 months or longer with follow-up (77,87). A relationship appears to exist between cumulative training dose and final efficacy. One study on iPad-based binocular game training reported that the visual acuity improvement in the binocular treatment group was significantly better than that in the control group at 4 weeks. However, this difference was no longer present after 8 weeks (71). Another 16-week study found that binocular iPad game training was not superior to traditional patching in terms of the magnitude of visual acuity improvement (53). Another study indicated no significant difference in efficacy between 4 and 8 weeks of binocular game training, suggesting a possible plateau phase after initial improvement (72). To mitigate dose-related bias, a few RCTs have incorporated equivalent training doses in their design, matching the cumulative duration of binocular training with that of occlusion therapy (54,69,70). Meanwhile, other studies have prioritized clinical feasibility and family acceptability, employing exploratory protocols with lower frequency and shorter periods (51,74,85). In terms of technological approaches, perceptual learning studies often adopt a pattern of high repetition and short cycles, emphasizing the relationship between early changes and long-term trends (43). In contrast, VR and gamified training place greater emphasis on sustained engagement, with training frequencies typically set at 3 to 7 days per week (57,71,75). These differences suggest that training dose is not a simple linear metric, and thus a comprehensive consideration of factors such as age, amblyopia type, and the specific neural pathway targeted is required. The optimal dosage range remains to be further defined within a unified evaluation framework.

Discussion

This narrative review synthesizes the current evidence on behavioral vision training for childhood amblyopia across a range of technical modalities. Several main findings emerge from the literature, each with implications for clinical practice and future research.

Overall strength and limitations of the evidence. The evidence base has expanded considerably since 2015, with the publication of several large, well-designed RCTs (e.g., CureSight, Luminopia One, Pediatric Eye Disease Investigator Group iPad game trials). These provide moderate-to-high certainty evidence for specific interventions in defined populations. However, many studies remain small, single-arm, or lack masked outcome assessment, introducing risk of bias. The heterogeneity in training protocols (task type, dose, frequency, duration) and outcome measures (different visual acuity charts, stereo tests) limits direct comparisons and precludes meta-analysis. Consequently, the overall certainty of evidence for behavioral vision training as a class of interventions is moderate at best.

Explaining mixed comparative effectiveness. The finding that some studies show behavioral training superior to patching, others show equivalence, and still others show inferiority is not necessarily contradictory. Discrepancies likely arise from differences in: (I) patient selection—younger children (4–6 years) may respond better to gamified interventions than older children (7–12 years); (II) control group intensity—studies using 2 hours of daily patching as control are more likely to show noninferiority than those using 6 hours; (III) training dose and duration—short-term (4–8 weeks) gains with binocular treatment may plateau or regress over longer follow-up; (IV) outcome timing—early improvements at 1 month do not always persist to 6 months. These factors should be considered when interpreting head-to-head comparisons.

Binocular function as a distinct therapeutic target. Across studies, improvements in stereopsis and CS are not consistently correlated with visual acuity gains. Some interventions (e.g., passive contrast-rebalanced videos) improve acuity without affecting stereoacuity, whereas others (e.g., dichoptic action games) show preferential effects on binocular function. This suggests that monocular and binocular deficits in amblyopia may be mediated by partially separable neural mechanisms. Optimal treatment may therefore require a combination of modalities targeting each domain, or sequential approaches that address monocular acuity first followed by binocular integration.

Clinical practice guidelines and regulatory status. As of 2026, only a limited number of behavioral vision training devices have received regulatory clearance (e.g., Luminopia One for refractive amblyopia in children aged 4–7 years). Most other dichoptic and gamified interventions remain classified as investigational or are used off-label. The AAPOS preferred practice pattern continues to recommend patching or atropine as standard care, while acknowledging that binocular therapies may be considered as adjuncts or in cases of poor adherence to conventional treatment (7). Clinicians should therefore inform caregivers about the experimental nature of most behavioral training modalities and discuss the uncertainty regarding long-term outcomes.

Adherence considerations. While behavioral training generally reports higher short-term adherence (80–98%) than traditional patching (often 50–70% in real-world settings), long-term adherence declines across all modalities. The novelty effect of gamified interventions may diminish over weeks to months. Home-based unsupervised training remains vulnerable to poor compliance despite device logs; supervised or remotely monitored sessions appear to improve adherence. Clinical implementation should incorporate strategies to sustain motivation, such as periodic in-clinic reinforcement, caregiver education, and adaptive difficulty progression.

Challenges in clinical application and future directions

Current challenges

Presently, one of the most significant challenges in the field of behavioral vision training is the lack of unified, standardized treatment protocols (36). Considerable heterogeneity exists across studies in terms of training tasks, duration, frequency, and platforms. For example, training tasks range from orientation discrimination and contrast detection to various binocular fusion games. Home-based training sessions range from 20 to 90 minutes per day, and total intervention periods range from 10 days to 6 months (34,85,92,93). This variability makes it difficult to directly compare study results and presents challenges to clinicians seeking to develop specific treatment plans. More critically, clear guidance is currently lacking on the means to formulate personalized training strategies based on amblyopia type (e.g., anisometropic versus strabismic), severity, and individual patient characteristics (e.g., baseline visual acuity, refractive status, and age). One study indicated that poorer baseline visual acuity and the presence of high astigmatism may be risk factors for a less favorable response to adjunctive perceptual learning (66). This finding underscores the importance of individualized treatment approaches.

The overall quality of available evidence is unsatisfactory, and data on long-term efficacy are generally insufficient. Although the number of supporting studies has increased, high-quality RCTs remain relatively sparse. Many studies are retrospective, employ single-group, pretest-posttest designs, or are quasi-experimental, entailing a high risk of bias. A lack of masking in outcome assessment is a particular concern in many of these studies (52,94). A Cochrane systematic review noted that the quality of evidence comparing binocular treatment to standard patching or atropine therapy for monocular amblyopia was generally low to moderate, underscoring the need for more rigorously designed trials (93). Furthermore, the majority of research is based on short follow-up periods, and thus the stability of improvements in visual acuity and visual function in the long-term remains to be validated. Although a few studies have reported efficacy persisting for up to 3 to 6 months posttreatment (87,95), follow-up data extending beyond 1 year or longer remain scarce. This limits the ability to fully assess long-term effectiveness and the risk of recurrence.

Treatment adherence, particularly in home-based training settings, poses another significant challenge. Although gamification and digital design aim to enhance engagement and accessibility, the lack of direct supervision often leads to fluctuating or poor adherence. One study focusing on older, patching-resistant children with amblyopia found that self-administered perceptual learning via video games at home did not consistently improve visual function. However, data analysis revealed a positive correlation between treatment adherence and certain visual outcomes, such as binocular summation effects (92). This suggests that in the absence of effective supervision and feedback mechanisms, the potential efficacy of home-based training may not be fully realized. A key question for research is how protocols can be designed such that they are both effective, practical for home implementation, and easy to monitor.

Knowledge regarding the heterogeneity of treatment efficacy across different patient subgroups and the potential associated risks remains insufficient. The effectiveness of behavioral vision training may vary across patients of different ages and with different types of amblyopia. For instance, some studies suggest that internet-based perceptual learning may be more effective for patients with refractive and strabismic amblyopia, but its efficacy for those with anisometropic amblyopia might not be superior to conventional treatment (52). Of particular note, a recent large-scale retrospective study has raised concerns regarding the safety of digital therapies. It indicated that daily treatment with such therapies may be associated with an increased risk of rapid axial growth in children, especially those who are hyperopic or premyopic (96). This finding underscores the necessity of carefully monitoring and evaluating treatment-induced changes in ocular biometric parameters and the active pursuit of functional improvement.

The neural mechanisms underlying the therapeutic effects of behavioral vision training have not been fully elucidated, which limits the optimization of training strategies. One study observed that after dichoptic video game training, patients showed significant improvements in stereopsis and visual evoked potential latency, a measure reflecting visual processing speed, without concurrent improvements in visual acuity (97). This suggests that behavioral training may preferentially enhance higher-order visual processing functions, with acuity improvements potentially lagging behind or being mediated by different mechanisms. Neuroimaging studies have demonstrated that perceptual learning can induce plastic changes in V1 function and that these changes correlate with clinical improvements in visual acuity and stereopsis (98,99). Electroencephalography studies have also confirmed that dichoptic attention task training can enhance attentional modulation in the V1 and the intraparietal sulcus while reducing interocular suppression (100). However, it is not yet clear how these neuroscientific findings can be translated into concrete guidance for optimizing training strategies, such as determining whether tasks targeting low-level visual features or higher-order cognitive functions are more effective; moreover, the means to optimally combine different training modules remains to be established.

Future directions

To address the challenges in this field and translate the potential of behavioral vision training into widespread and reliable clinical benefits, future research should follow a translational medicine pathway. Systematic exploration is needed across multiple dimensions, which includes strengthening the evidence base, deepening mechanistic understanding, optimizing technology, and integrating strategies. A key task is the promotion of high-quality clinical research and the development of a standardized framework. Although the preliminary evidence is encouraging, the studies from which it has been derived have limitations in sample size, methodological rigor, and follow-up duration. Therefore, there is a need for large-scale, multicenter, long-term RCTs, complemented by real-world studies. This type of research would provide high-level evidence regarding efficacy, long-term stability, and the risk of recurrence. These studies should pay particular attention to the long-term stability of treatment effects and recurrence rates and should include comprehensive assessments of stereopsis, CS, and quality of life. Moreover, it is crucial to develop consensus-based standardized treatment protocols and clinical practice guidelines in collaboration with professional societies, clinical experts, and methodologists. This framework should define the core training elements, such as task type, dose, and frequency, as well as the uniform criteria for efficacy evaluation. Doing so would provide a structure for standardized application across different clinical settings and a solid foundation for the development of personalized treatment plans.

Future research should focus on elucidating the mechanisms of neural plasticity and developing precision-based, intelligent treatment models. First, efforts should be directed toward developing home-based training systems that are low-cost and highly accessible and that feature child-friendly interfaces. Integrating remote monitoring and feedback capabilities into these systems would help optimize treatment adherence and accessibility. Second, the development of adaptive training platforms that integrate artificial intelligence and machine learning algorithms is a key priority. Such systems could dynamically adjust task difficulty, contrast, and interocular balance parameters based on a patient’s real-time performance data, such as reaction time and accuracy. This would enable truly personalized prescriptions. Furthermore, evaluating the feasibility of combining behavioral training with neuromodulation techniques, such as transcranial electrical stimulation, may open novel avenues for enhancing brain plasticity and treating amblyopia that is resistant to conventional therapies.

Ultimately, the focus of research should return to the comprehensive optimization and safe expansion of clinical practice. This requires that future work not only concentrate on improvements in visual acuity but also establish a multidimensional evaluation system. Such a system should encompass stereopsis, CS, visual-information-processing efficiency, and patient quality of life. Alongside the active promotion of digital therapeutics, it is essential to implement routine, long-term safety monitoring protocols. Regular follow-up of refractive status and axial length is particularly important for scientifically evaluating and managing potential biomechanical risks. Furthermore, behavioral vision training, traditional occlusion, and optical correction should not be viewed as mutually exclusive alternatives but rather as complementary and synergistic approaches. A key direction for future research is to determine how best to integrate behavioral training into the existing treatment ladder for amblyopia. This includes determining the optimal timing and mode of intervention at different stages—such as initial treatment and consolidation—or for treatment-resistant cases. The goal should be the development of efficient, personalized, and safe integrated treatment strategies that maximize patients’ long-term visual rehabilitation outcomes.

Conclusions

Behavioral vision training is an active intervention grounded in the principles of neuroplasticity. Its role in the treatment of childhood amblyopia is no longer limited to being an adjunct to traditional occlusion therapy. As research has advanced, this form of training has gradually been integrated into the comprehensive framework for the management of amblyopia and has demonstrated independent and stable clinical value. The training modalities of behavioral vision training encompass a range of technological approaches, including perceptual learning, dichoptic stimulation, digital game-based training, and VR. This variety reflects the ongoing diversification of intervention strategies. Studies on behavioral vision training suggest it not only improves BCVA in the amblyopic eye but also supports holistic rehabilitation of the binocular visual system. At the level of binocular function, parameters such as stereopsis, CS, and accommodative facility are considered important therapeutic targets. Compared to patching therapy, behavioral training typically offers greater interactivity and situational engagement. This characteristic aligns well with the cognitive and behavioral profiles of children. Clinical observations indicate that the engaging nature of the training format may encourage participation from children, offering a novel practical approach to sustaining long-term intervention.

There is also substantial heterogeneity in treatment efficacy reported across different studies. Outcomes appear to be influenced by multiple factors, including the specific training paradigm, the child’s developmental status, and the parameters of the treatment protocol. Regarding mechanisms of action and therapeutic goals, behavioral vision training and occlusion therapy are not mutually exclusive alternatives but rather exhibit a clear complementary relationship. Integrating these two approaches, either sequentially or concurrently, based on the individual child’s condition, may help optimize the trajectory of visual function development while also promoting adherence. Future research should continue to focus on establishing standardized protocols, developing personalized strategies, conducting comprehensive safety evaluations, and exploring multimodal synergies. These efforts are essential to advancing the treatment of amblyopia toward an active, precise model of neural functional remodeling.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the Narrative Review reporting checklist. Available at https://tp.amegroups.com/article/view/10.21037/tp-2026-0464/rc

Peer Review File: Available at https://tp.amegroups.com/article/view/10.21037/tp-2026-0464/prf

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tp.amegroups.com/article/view/10.21037/tp-2026-0464/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Hu B, Liu Z, Zhao J, et al. The Global Prevalence of Amblyopia in Children: A Systematic Review and Meta-Analysis. Front Pediatr 2022;10:819998. [Crossref] [PubMed]
2. Fu Z, Hong H, Su Z, et al. Global prevalence of amblyopia and disease burden projections through 2040: a systematic review and meta-analysis. Br J Ophthalmol 2020;104:1164-70. [Crossref] [PubMed]
3. Li YP, Zhou MW, Forster SH, et al. Prevalence of amblyopia among preschool children in central south China. Int J Ophthalmol 2019;12:820-5. [Crossref] [PubMed]
4. Ming X, Huang G, Chen X, et al. A Systematic Review and Meta-Analysis of Perceptual Learning and Video Game Training for Adults with Monocular Amblyopia. Ophthalmol Ther 2025;14:857-81. [Crossref] [PubMed]
5. Li Y, Sun H, Zhu X, et al. Efficacy of interventions for amblyopia: a systematic review and network meta-analysis. BMC Ophthalmol 2020;20:203. [Crossref] [PubMed]
6. Zhang W, Shi XF, Li XT. Further improvement in the prevention and management of amblyopia in China. Chinese Journal of Ophthalmology 2025;61:1-6.
7. Cruz OA, Repka MX, Hercinovic A, et al. Amblyopia Preferred Practice Pattern. Ophthalmology 2023;130:136-78. [Crossref] [PubMed]
8. Chen DM, Han S, Summers A, et al. Interventions for improving adherence to amblyopia treatments in children. Cochrane Database Syst Rev 2025;7:CD015820. [Crossref] [PubMed]
9. Thompson B, Concetta Morrone M, Bex P, et al. Harnessing brain plasticity to improve binocular vision in amblyopia: An evidence-based update. Eur J Ophthalmol 2024;34:901-12. [Crossref] [PubMed]
10. Chaturvedi I, Jamil R, Sharma P. Binocular vision therapy for the treatment of Amblyopia-A review. Indian J Ophthalmol 2023;71:1797-803. [Crossref] [PubMed]
11. He Y, Feng L, Zhou Y, et al. Characteristics and predictive factors of visual function improvements after monocular perceptual learning in amblyopia. Heliyon 2023;9:e17281. [Crossref] [PubMed]
12. Yeritsyan A, Surve AV, Ayinde B, et al. Efficacy of Amblyopia Treatments in Children Up to Seven Years Old: A Systematic Review. Cureus 2024;16:e56705. [Crossref] [PubMed]
13. Repka MX. Amblyopia Outcomes Through Clinical Trials and Practice Measurement: Room for Improvement: The LXXVII Edward Jackson Memorial Lecture. Am J Ophthalmol 2020;219:A1-26. [Crossref] [PubMed]
14. Sen S, Singh P, Saxena R. Management of amblyopia in pediatric patients: Current insights. Eye (Lond) 2022;36:44-56. [Crossref] [PubMed]
15. Şahin Karamert S, Atalay HT, Özdek Ş. Strabismus in Retinopathy of Prematurity: Risk Factors and the Effect of Macular Ectopia. Turk J Ophthalmol 2023;53:241-6. [Crossref] [PubMed]
16. Li Y, Zheng G, Wen B, et al. Altered spontaneous brain activity in children with deprivation amblyopia: a resting-state functional magnetic resonance imaging study. Eur J Med Res 2025;30:31. [Crossref] [PubMed]
17. Magli A, Esposito Veneruso P, Rinaldi M, et al. Long-term effects of early/late-onset visual deprivation on macular and retinal nerve fibers layer structure: A pilot study. PLoS One 2023;18:e0283423. [Crossref] [PubMed]
18. Hu J, Chen J, Ku Y, et al. Reduced interocular suppression after inverse patching in anisometropic amblyopia. Front Neurosci 2023;17:1280436. [Crossref] [PubMed]
19. AlTurki H, Amir A, AlShamlan F, et al. A case of reverse amblyopia in a myopic anisometropic patient: an atypical presentation. J Surg Case Rep 2024;2024:rjae320. [Crossref] [PubMed]
20. Gorzek R, Trachtenberg JT. A Critical Look at Critical Periods. Annu Rev Vis Sci 2025;11:175-92. [Crossref] [PubMed]
21. Sahyoun GM, Do TD, Anqueira-Gonzàlez A, et al. Peripuberty Is a Sensitive Period for Prefrontal Parvalbumin Interneuron Activity to Impact Adult Cognitive Flexibility. Dev Neurosci 2025;47:127-38. [Crossref] [PubMed]
22. Fakheir Y, Khalil R. The effects of abnormal visual experience on neurodevelopmental disorders. Dev Psychobiol 2023;65:e22408. [Crossref] [PubMed]
23. Levi DM. Amblyopia. Handb Clin Neurol 2021;178:13-30. [Crossref] [PubMed]
24. Echavarri-Leet MP, Resnick HH, Bowen DA, et al. Spontaneous recovery from amblyopia following fellow eye vision loss: a systematic review and narrative synthesis. J AAPOS 2024;28:103971. [Crossref] [PubMed]
25. Hess RF, Mansouri B, Thompson B. A binocular approach to treating amblyopia: antisuppression therapy. Optom Vis Sci 2010;87:697-704. [Crossref] [PubMed]
26. Qin Y, Ahmadlou M, Suhai S, et al. Thalamic regulation of ocular dominance plasticity in adult visual cortex. Elife 2023;12:RP88124. [Crossref] [PubMed]
27. Jamal YA, Dilks DD. Rapid topographic reorganization in adult human primary visual cortex (V1) during noninvasive and reversible deprivation. Proc Natl Acad Sci U S A 2020;117:11059-67. [Crossref] [PubMed]
28. van Kerkoerle T, Marik SA, Meyer Zum Alten Borgloh S, et al. Axonal plasticity associated with perceptual learning in adult macaque primary visual cortex. Proc Natl Acad Sci U S A 2018;115:10464-9. [Crossref] [PubMed]
29. Carvalho J, Invernizzi A, Martins J, et al. Local neuroplasticity in adult glaucomatous visual cortex. Sci Rep 2022;12:21981. [Crossref] [PubMed]
30. Zhang X, Tang H, Li S, et al. Inhibition of Cdk5 in PV Neurons Reactivates Experience-Dependent Plasticity in Adult Visual Cortex. Int J Mol Sci 2021;23:186. [Crossref] [PubMed]
31. Xin W, Kaneko M, Roth RH, et al. Oligodendrocytes and myelin limit neuronal plasticity in visual cortex. Nature 2024;633:856-63. [Crossref] [PubMed]
32. Acquafredda M, Kurzawski JW, Biagi L, et al. The pulvinar regulates plasticity in human visual cortex. Sci Adv 2025;11:eadw9988. [Crossref] [PubMed]
33. Liu Y, Zhang J, Jiang Z, et al. Organization of corticocortical and thalamocortical top-down inputs in the primary visual cortex. Nat Commun 2024;15:4495. [Crossref] [PubMed]
34. Tsaousis KT, Mousteris G, Diakonis V, et al. Current Developments in the Management of Amblyopia with the Use of Perceptual Learning Techniques. Medicina (Kaunas) 2023;60:48. [Crossref] [PubMed]
35. Lin W, He Z, Zhou S, et al. Monocular Contrast Sensitivity Visual Perceptual Learning Rebalances Adult Amblyopes' Two Eyes. Invest Ophthalmol Vis Sci 2025;66:25. [Crossref] [PubMed]
36. Hernández-Andrés R, Serrano MÁ, Alacreu-Crespo A, et al. Randomised trial of three treatments for amblyopia: Vision therapy and patching, perceptual learning and patching alone. Ophthalmic Physiol Opt 2025;45:31-42. [Crossref] [PubMed]
37. Barollo M, Contemori G, Battaglini L, et al. Perceptual learning improves contrast sensitivity, visual acuity, and foveal crowding in amblyopia. Restor Neurol Neurosci 2017;35:483-96. [Crossref] [PubMed]
38. Jiang SQ, Chen YR, Liu XY, et al. Contour integration deficits at high spatial frequencies in children treated for anisometropic amblyopia. Front Neurosci 2023;17:1160853. [Crossref] [PubMed]
39. Chen YR, Jiang SQ, Liu XY, et al. Temporal Contour Integration Deficits in Children With Amblyopia. Invest Ophthalmol Vis Sci 2025;66:27. [Crossref] [PubMed]
40. Xu K, Luo W, Ye H, et al. Home-based perceptual learning augments high-frequency contrast sensitivity and stereopsis after esotropia surgery: A retrospective cohort study. J Optom 2026;19:100589. [Crossref] [PubMed]
41. Zhou Y, Ye Q, Qiu X, et al. Extending Treatment Duration in Perceptual Learning for Amblyopia. Ophthalmol Sci 2026;6:101005. [Crossref] [PubMed]
42. Eisen-Enosh A, Farah N, Polat U, et al. Perceptual learning based on a temporal stimulus enhances visual function in adult amblyopic subjects. Sci Rep 2023;13:7643. [Crossref] [PubMed]
43. He Y, Xu Z, Feng L, et al. Personalized Rehabilitation for Residual Deficits: Tailoring Perceptual Learning for Improved Visual Function in Meridional Amblyopia. Ophthalmol Sci 2025;5:100736. [Crossref] [PubMed]
44. do Amaral Júnior FL, Rodrigues TA, da Silva NLT, et al. Comparative Neuroplasticity in Frontal- and Lateral-Eyed Mammals With Induced-Binocular Vision Dysfunction: Insights From Monocular Deprivation Models. Eur J Neurosci 2025;62:e70179. [Crossref] [PubMed]
45. Iwata Y, Handa T, Ishikawa H. Comparison of Amblyopia Treatment Effect with Dichoptic Method Using Polarizing Film and Occlusion Therapy Using an Eye Patch. Children (Basel) 2022;9:1285. [Crossref] [PubMed]
46. Quagraine IM, Murray J, Cakir GB, et al. Evaluating Eye Tracking During Dichoptic Video Viewing With Varied Fellow Eye Contrasts in Amblyopia. Invest Ophthalmol Vis Sci 2024;65:11. [Crossref] [PubMed]
47. Ansari AAA, Sharma P, Khateeb EW, et al. Effectiveness of binocular vision therapy in managing patients with intermittent exotropia. Indian J Ophthalmol 2025;73:1422-5. [Crossref] [PubMed]
48. Liu XY, Zhang YW, Gao F, et al. Dichoptic Perceptual Training in Children With Amblyopia With or Without Patching History. Invest Ophthalmol Vis Sci 2021;62:4. [Crossref] [PubMed]
49. Lan FF, Zhao WX, Gan L, et al. Repeated visual perceptual micro-training enhances neural microplasticity in children with ametropic amblyopia: A clinical observation. Photodiagnosis Photodyn Ther 2026;57:105311. [Crossref] [PubMed]
50. Asare AK, Ho CS, Im HY, et al. Evaluation of motion perception and binocular vision following dichoptic treatment for amblyopia. Vision Res 2026;240:108745. [Crossref] [PubMed]
51. Zhu W, Gu S, Li J, et al. Transformative Gamified Binocular Therapy for Unilateral Amblyopia in Young Children: Pilot Prospective Efficacy and Safety Study. JMIR Serious Games 2025;13:e63384. [Crossref] [PubMed]
52. Schmucker C, Thörel E, Flatscher-Thöni M, et al. Computer-Assisted Visual Training in Children and Adolescents with Developmental Visual Disorders. Dtsch Arztebl Int 2023;120:747-53. [Crossref] [PubMed]
53. Holmes JM, Manh VM, Lazar EL, et al. Effect of a Binocular iPad Game vs Part-time Patching in Children Aged 5 to 12 Years With Amblyopia: A Randomized Clinical Trial. JAMA Ophthalmol 2016;134:1391-400. [Crossref] [PubMed]
54. Chen Y, Chen Y, Han X, et al. Comparative effectiveness of gamified binocular treatment versus conventional patching for amblyopia: a randomized clinical trial. Front Med (Lausanne) 2025;12:1560203. [Crossref] [PubMed]
55. Coco-Martin MB, Piñero DP, Leal-Vega L, et al. The Potential of Virtual Reality for Inducing Neuroplasticity in Children with Amblyopia. J Ophthalmol 2020;2020:7067846. [Crossref] [PubMed]
56. Nagino K, Okumura Y, Hirota M, et al. Virtual Reality-Based Program for Pediatric Patients With Amblyopia: Protocol for a Multicenter, Randomized, Open-Label, Two-Arm Study. JMIR Res Protoc 2026;15:e85194. [Crossref] [PubMed]
57. Xiao S, Angjeli E, Wu HC, et al. Randomized Controlled Trial of a Dichoptic Digital Therapeutic for Amblyopia. Ophthalmology 2022;129:77-85. [Crossref] [PubMed]
58. Gao Y, Zhou Y, He Q, et al. Improving Adult Vision Through Pathway-Specific Training in Augmented Reality. Adv Sci (Weinh) 2025;12:e2415877. [Crossref] [PubMed]
59. Hernández-Rodríguez CJ, Ferrer-Soldevila P, Artola-Roig A, et al. Rehabilitation of amblyopia using a digital platform for visual training combined with patching in children: a prospective study. Graefes Arch Clin Exp Ophthalmol 2024;262:3007-20. [Crossref] [PubMed]
60. Piñero DP, Gil-Casas A, Hurtado-Ceña FJ, et al. Visual Performance of Children with Amblyopia after 6 Weeks of Home-Based Dichoptic Visual Training. Children (Basel) 2024;11:1007. [Crossref] [PubMed]
61. Liu B, Fan Y, Xu M, et al. Effectiveness of a Gamified Mobile App in Enhancing Treatment Adherence for Children With Amblyopia: Explorative Study. JMIR Serious Games 2025;13:e60309. [Crossref] [PubMed]
62. Uttamapinan S, Pukrushpan P, Honglertnapakul W. Effectiveness of the smartphone application in increasing compliance with occlusion therapy in children with amblyopia: a randomized controlled trial. Strabismus 2024;32:73-80. [Crossref] [PubMed]
63. Krungkraipetch L, Supajitgulchai D, Assawaboonyadech A, et al. Short- and long-term outcomes of binocular gaming versus patching in childhood amblyopia: a systematic review and meta-analysis. Eur J Pediatr 2026;185:73. [Crossref] [PubMed]
64. Hernández-Rodríguez CJ, Fukumitsu H, Ruiz-Fortes P, et al. Efficacy of Perceptual Learning-Based Vision Training as an Adjuvant to Occlusion Therapy in the Management of Amblyopia: A Pilot Study. Vision (Basel) 2021;5:15. [Crossref] [PubMed]
65. Lan FF, Zhao WX, Gan L. Evaluation of visual plasticity in patients with refractive amblyopia treated using a visual perceptual learning system. Technol Health Care 2024;32:327-33. [Crossref] [PubMed]
66. Huang HC, Cho WH, Fang PC, et al. Add-on perceptual learning on refractive amblyopia in children. Int J Ophthalmol 2024;17:1850-6. [Crossref] [PubMed]
67. Khorrami-Nejad M, Akbari MR, Abdulhussein R, et al. Comparison of Cambridge vision stimulator (CAM) therapy with passive occlusion therapy in the management of unilateral amblyopia; a randomized clinical trial. Strabismus 2024;32:123-38. [Crossref] [PubMed]
68. Huang HM, Hsiao YT, Chen YH, et al. Comparison of Virtual Reality-Assisted Visual Training with Conventional Strategies in the Treatment of Bilateral Refractive Amblyopia. Children (Basel) 2025;12:447. [Crossref] [PubMed]
69. Wygnanski-Jaffe T, Kushner BJ, Moshkovitz A, et al. An Eye-Tracking-Based Dichoptic Home Treatment for Amblyopia: A Multicenter Randomized Clinical Trial. Ophthalmology 2023;130:274-85. [Crossref] [PubMed]
70. Wygnanski-Jaffe T, Kushner BJ, Moshkovitz A, et al. High-Adherence Dichoptic Treatment Versus Patching in Anisometropic and Small Angle Strabismus Amblyopia: A Randomized Controlled Trial. Am J Ophthalmol 2025;269:293-302. [Crossref] [PubMed]
71. Manny RE, Holmes JM, Kraker RT, et al. A Randomized Trial of Binocular Dig Rush Game Treatment for Amblyopia in Children Aged 4 to 6 Years. Optom Vis Sci 2022;99:213-27. [Crossref] [PubMed]
72. Pediatric Eye Disease Investigator Group. A Randomized Trial of Binocular Dig Rush Game Treatment for Amblyopia in Children Aged 7 to 12 Years. Ophthalmology 2019;126:456-66.
73. Jost RM, Birch EE, Wang YZ, et al. Patch-free streaming contrast-rebalanced dichoptic cartoons versus patching for treatment of amblyopia in children aged 3 to 5 years: a pilot, randomized clinical trial. J AAPOS 2024;28:103991. [Crossref] [PubMed]
74. Jost RM, Hudgins LA, Dao LM, et al. Randomized clinical trial of streaming dichoptic movies versus patching for treatment of amblyopia in children aged 3 to 7 years. Sci Rep 2022;12:4157. [Crossref] [PubMed]
75. Kadhum A, Tan ETC, Fronius M, et al. Supervised dichoptic gaming versus monitored occlusion therapy for childhood amblyopia: Effectiveness and efficiency. Acta Ophthalmol 2024;102:38-48. [Crossref] [PubMed]
76. Poltavski D, Adams RJ, Biberdorf D, et al. Effectiveness of a Novel Video Game Platform in the Treatment of Pediatric Amblyopia. J Pediatr Ophthalmol Strabismus 2024;61:20-9. [Crossref] [PubMed]
77. Liang J, Pang S, Yan L, et al. Efficacy of binocular vision training and Fresnel press-on prism on children with esotropia and amblyopia. Int Ophthalmol 2023;43:583-8. [Crossref] [PubMed]
78. Roy S, Saxena R, Dhiman R, et al. Comparison of Dichoptic Therapy Versus Occlusion Therapy in Children With Anisometropic Amblyopia: A Prospective Randomized Study. J Pediatr Ophthalmol Strabismus 2023;60:210-7. [Crossref] [PubMed]
79. Dahlmann-Noor AH, Greenwood JA, Skilton A, et al. Feasibility of a new 'balanced binocular viewing' treatment for unilateral amblyopia in children aged 3-8 years (BALANCE): results of a phase 2a randomised controlled feasibility trial. BMJ Open 2024;14:e082472. [Crossref] [PubMed]
80. Leal-Vega L, Coco-Martín M. NEIVATECH pilot study: immersive virtual reality training in older amblyopic children with non-compliance or non-response to patching. Sci Rep 2024;14:28062. [Crossref] [PubMed]
81. Herbison N, MacKeith D, Vivian A, et al. Randomised controlled trial of video clips and interactive games to improve vision in children with amblyopia using the I-BiT system. Br J Ophthalmol 2016;100:1511-6. [Crossref] [PubMed]
82. Xiao S, Gaier ED, Wu HC, et al. Digital therapeutic improves visual acuity and encourages high adherence in amblyopic children in open-label pilot study J AAPOS 2021;25:87.
83. Rajavi Z, Soltani A, Vakili A, et al. Virtual Reality Game Playing in Amblyopia Therapy: A Randomized Clinical Trial. J Pediatr Ophthalmol Strabismus 2021;58:154-60. [Crossref] [PubMed]
84. Suwal R, Dev MK, Khatri B, et al. Impact of active vision therapy compared to conventional patching therapy on visual acuity and stereoacuity in children with amblyopia. J Optom 2024;17:100484. [Crossref] [PubMed]
85. Mo Y, Chen P, Hou M, et al. Serious Games Integrating Perceptual Learning and Stereopsis Training in Children With Amblyopia: Single-Arm Pre-Post Feasibility Study. JMIR Serious Games 2025;13:e77402. [Crossref] [PubMed]
86. Portela-Camino JA, Martín-González S, Ruiz-Alcocer J, et al. An Evaluation of the Agreement Between a Computerized Stereoscopic Game Test and the TNO Stereoacuity Test. Clin Optom (Auckl) 2021;13:181-90. [Crossref] [PubMed]
87. Milla M, Molina-Martín A, Piñero DP. Long-Term Efficacy of the Combination of Active Vision Therapy and Occlusion in Children with Strabismic and Anisometropic Amblyopia. Children (Basel) 2022;9:1012. [Crossref] [PubMed]
88. Asensio-Jurado L, Argilés M, Vinuela-Navarro V, et al. Efficacy of patching combined with action video games in amblyopic children aged 4-10 years: A randomised clinical trial. Ophthalmic Physiol Opt 2025;45:1389-98. [Crossref] [PubMed]
89. Singh A, Sharma P, Saxena R. Evaluation of the Role of Monocular Video Game Play as an Adjuvant to Occlusion Therapy in the Management of Anisometropic Amblyopia. J Pediatr Ophthalmol Strabismus 2017;54:244-9. [Crossref] [PubMed]
90. Li L, Xue H, Lai T, et al. Comparison of compliance among patients with pediatric amblyopia undergoing virtual reality-based and traditional patching method training. Front Public Health 2022;10:1037412. [Crossref] [PubMed]
91. Kadhum A, Tan ETC, Levi DM, et al. Barriers to successful dichoptic treatment for amblyopia in young children. Graefes Arch Clin Exp Ophthalmol 2021;259:3149-57. [Crossref] [PubMed]
92. Lee YH, Maniglia M, Velez F, et al. Short-term Perceptual Learning Game Does Not Improve Patching-Resistant Amblyopia in Older Children. J Pediatr Ophthalmol Strabismus 2020;57:176-84. [Crossref] [PubMed]
93. Tailor V, Ludden S, Bossi M, et al. Binocular versus standard occlusion or blurring treatment for unilateral amblyopia in children aged three to eight years. Cochrane Database Syst Rev 2022;2:CD011347. [Crossref] [PubMed]
94. Hernández-Rodríguez CJ, Piñero DP. Active Vision Therapy for Anisometropic Amblyopia in Children: A Systematic Review. J Ophthalmol 2020;2020:4282316. [Crossref] [PubMed]
95. Hsieh YC, Liao WL, Tsai YY, et al. Efficacy of vision therapy for unilateral refractive amblyopia in children aged 7-10 years. BMC Ophthalmol 2022;22:44. [Crossref] [PubMed]
96. Yao Y, He Y, Wen Y, et al. Factual Evidence on Digital Therapeutics in Pediatric Amblyopia: Insights into Rapid Axial Elongation Risk. Ophthalmology 2025;132:661-70. [Crossref] [PubMed]
97. Blavakis E, Spaho J, Chatzea M, et al. Dichoptic Game Training in Strabismic Amblyopia Improves the Visual Evoked Response. Cureus 2023;15:e45395. [Crossref] [PubMed]
98. Hou C, Zhou Z, Uner IJ, et al. Visual Cortical Function Changes After Perceptual Learning with Dichoptic Attention Tasks in Adults with Amblyopia: A Case Study Evaluated Using fMRI. Brain Sci 2024;14:1148. [Crossref] [PubMed]
99. Consorti A, Sansevero G, Torelli C, et al. Visual Perceptual Learning Induces Long-Lasting Recovery of Visual Acuity, Visual Depth Perception Abilities and Binocular Matching in Adult Amblyopic Rats. Front Cell Neurosci 2022;16:840708. [Crossref] [PubMed]
100. Hou C, Nicholas SC. Perceptual learning with dichoptic attention tasks improves attentional modulation in V1 and IPS and reduces interocular suppression in human amblyopia. Sci Rep 2022;12:9660. [Crossref] [PubMed]

(English Language Editor: J. Gray)