Lazy eye treatment research: what the evidence says about dichoptic therapy

Key takeaways

• Standard treatments are the established options. Correcting the underlying problem with glasses, contacts or surgery, patching the stronger eye, and atropine drops are the treatments the National Eye Institute lists [1].
• There is a scientific reason to try dichoptic treatment. Laboratory work suggests that suppression of the weaker eye's signal is a key barrier to improvement, and that presenting images separately to each eye can reduce it [12][11].
• Clinical results are mixed. A dichoptic headset therapy beat glasses alone in children aged 4 to 7 [8], and dichoptic movies matched patching over 2 weeks [7]. But the largest trial of an iPad game did not show it was as good as patching [2]. In teenagers it was not better than patching, and possibly worse [6]. In previously treated children aged 7 to 12 it gave no extra benefit over glasses [3].
• Newer app studies report encouraging results, but so far in small pilot studies without a comparison group of children with amblyopia [24].
• A Cochrane review found only one eligible trial for children aged 3 to 8 and judged binocular and patching results likely comparable at 16 weeks, with limited evidence overall [9].
• Adherence is the recurring problem. In several trials, only a minority of children completed most of the prescribed play [2][6][9].
• One dichoptic product is FDA-authorized (Luminopia One, for ages 4 to 7, on a headset, by prescription) [14].
• ZyncView has not been tested in a published trial and is not FDA-authorized. This page describes other products' studies. Their results may not carry over.

Background and how common it is

The National Eye Institute (NEI) describes amblyopia as a type of poor vision that usually happens in one eye, less commonly in both, where the brain can't recognize the sight from that eye. It says up to 3 out of 100 children have it and that it is the most common cause of vision loss in children [1]. Underlying eye conditions that can lead to it include refractive errors, misaligned eyes (strabismus) and cataracts, although a specific cause is often not found [1].

A 2020 systematic review and meta-analysis pooled 60 studies with 1,859,327 participants and estimated the worldwide prevalence at 1.44% (95% confidence interval 1.17% to 1.78%). Estimates were higher in Europe (2.90%) and North America (2.41%) than in Asia (1.09%) and Africa (0.72%), and the highest prevalence (3.29%) was found in people over 20 years old. The authors estimated 99.2 million people with amblyopia in 2019, rising to 221.9 million by 2040, and called for screening, treatment and related public health strategies [10]. Numbers differ between sources because studies define amblyopia differently and sample different populations, so treat any single figure as an estimate.

The science: suppression and plasticity

Amblyopia was once considered a structurally "monocular" problem, but a 2014 review by Hess, Thompson and Baker explains that the capacity for binocular vision is present in many people with amblyopia. Better ways of measuring suppression (the brain turning down the signal of the weaker eye) have shown how it relates to the visual deficits people experience. The review adds that new treatments target these suppressive interactions and, on the basis of initial data, appear to improve both binocular and monocular visual function, even in adults [12].

A laboratory study by Li and colleagues in 2013 tested the idea directly in adults with amblyopia, who usually do not respond to conventional patching. Their results indicated that presenting stimuli separately to the two eyes (dichoptic presentation) reduced suppression and produced greater plasticity than forced use of the amblyopic eye alone. They concluded that suppression is a key gating mechanism that prevents the amblyopic brain from learning to see [11].

A separate line of work, perceptual learning, looks at whether practising a visual task improves vision. A 2009 mini-review by Levi and Li concluded that practising a visual task results in a long-lasting improvement in performance in an amblyopic eye, and asked how such training might be made more effective and efficient [13].

What these studies do and do not show. They describe a plausible mechanism and promising laboratory effects. They are not, by themselves, proof that a particular game improves everyday vision in a particular patient. That is what clinical trials are for.

Standard treatments and their limits

The NEI lists three main approaches: correcting the underlying problem with glasses, contact lenses or surgery; wearing an eye patch over the stronger eye; and putting special (atropine) drops in the stronger eye. It stresses that it is important to start early ("the sooner the better") and notes that treatment is usually less effective in adults than in children [1].

The authors of a Cochrane review describe the limits of current treatment: less than two-thirds of children achieve good acuity (0.20 logMAR or better) in the amblyopic eye, improvement in stereopsis (depth perception) is limited, and adherence to treatment is poor. That is the motivation for testing new approaches such as dichoptic movies and games, which aim to balance the input from each eye and may be more engaging [9].

Seven trials at a glance

One "line" on an eye chart is 5 letters, or 0.1 logMAR. PEDIG is the Pediatric Eye Disease Investigator Group. Trials differ in ages, prior treatment, comparison treatment and length, so their numbers cannot simply be added together.

The trials, one by one

Early and small studies (2015 and 2016)

Birch and colleagues gave 50 preschool children (3 to under 7 years) binocular iPad games for at least 4 hours a week for 4 weeks. In the 45 children who played the binocular games, average acuity improved from 0.43 to 0.34 logMAR, while the 5 children given sham games did not improve significantly; two-thirds of the game group also patched at other times of day. The authors concluded binocular play was more effective than sham play in this age group [4]. The comparison group was very small, so this study raised the question rather than answering it.

Kelly and colleagues randomised 28 children (4.6 to 9.5 years) in a crossover design to a binocular game or patching for 2 weeks. At 2 weeks improvement was greater with the game (about 1.5 lines) than with patching (about 0.7 lines) [5]. The study was small and very short, but together with the previous one it justified larger trials.

The large PEDIG trials (2016 to 2019)

ATS18 enrolled 385 children aged 5 to under 13 with amblyopia caused by strabismus, anisometropia or both (acuity 20/40 to 20/200, average about 20/63). They were randomised to 16 weeks of a binocular iPad game for 1 hour a day or patching for 2 hours a day. Average improvement was 1.05 lines with the game and 1.35 lines with patching. The difference of 0.31 lines had an upper confidence limit of 0.53 lines, above the 0.5-line limit set in advance for calling the game "not worse" than patching, so non-inferiority was not shown. Only 39 of 176 children (22.2%) in the game group completed more than 75% of the prescribed treatment [2].

ATS19 asked the same question in 100 teenagers aged 13 to under 17. Acuity improved by 3.5 letters with the game and 6.5 letters with patching. After adjustment the difference was 2.7 letters (95% CI -5.7 to 0.3; P = 0.082), or 0.5 lines, favouring patching. Only 13% of the game group completed more than 75% of the treatment. The authors concluded that the game was not found to be better than patching and was possibly worse, but noted it remains unclear whether the small response was due to poor adherence or to a lack of treatment effect [6].

ATS20 tested the Dig Rush game in 138 children aged 7 to 12, 96% of whom had already had treatment other than glasses. Children were randomised to 8 weeks of the game (1 hour a day, 5 days a week) plus glasses, or to continued glasses alone. At 4 weeks acuity improved by 1.3 letters with the game and 1.7 letters with glasses alone (adjusted difference -0.3 letters, 95% CI -2.2 to 1.5), and there was no difference at 8 weeks [3].

Newer trials (2022)

Jost and colleagues randomised 65 children aged 3 to 7 to streaming dichoptic movies at home on a handheld 3D-enabled console (3 movies a week) or patching (14 hours a week). At the 2-week primary outcome, the movies improved acuity similarly to patching, and children in both groups improved further with up to 6 weeks of movie treatment. The authors describe movie treatment as "an additional treatment option" [7]. The main comparison was only 2 weeks long, so it says little about longer-term results.

Xiao and colleagues ran a phase 3 trial of a dichoptic digital therapeutic (the product later authorized by the FDA, see below) in 105 children aged 4 to 7, at 21 sites in the United States. Children in the treatment group used it at home for 1 hour a day, 6 days a week and wore glasses full-time; the comparison group wore glasses full-time alone. At 12 weeks acuity improved by 1.8 lines in the treatment group and 0.8 lines in the comparison group. The 1.0-line difference was statistically significant (P = 0.0011) and the study was stopped early for success at a planned interim analysis, with no serious adverse events [8]. Two things are worth keeping in mind: the comparison was glasses alone rather than patching, and the trial evaluated the developer's own product.

Newer app-based studies

Some newer studies look beyond eye-chart acuity. A 2025 pilot study from Turkey tested a gamified dichoptic app in 14 children aged 7 to 13 with anisometropic amblyopia. The children played for 20 minutes a day for 6 months. The comparison group was 15 typically developing children who had no intervention [24].

In the amblyopia group, acuity in the weaker eye improved from 0.23 to 0.08 logMAR on average, which is about 1.5 lines. Depth perception (stereopsis) improved from about 494 to 61 arcsec. The researchers also reported faster and steadier eye movements, and better scores on tests of motor skill and balance [24].

These are encouraging numbers, but the design limits what they show:

• It was a small pilot and not a randomised trial. The paper lists trial registration as "not applicable".
• The comparison children did not have amblyopia and had no treatment. So the study cannot separate the effect of the app from natural change, or from other treatment such as glasses. The abstract does not say whether the children used other treatments.
• Starting acuity averaged 0.23 logMAR (about 20/34), so many children had a fairly small deficit to close.
• Balance was measured with a posture test. Falls were not measured.

Pilot studies like this are useful for suggesting ideas to test. To show whether such an app adds benefit, a randomised trial with a comparison group of children who also have amblyopia would be needed [24].

The Cochrane systematic review

A Cochrane review (updated in 2022) looked for randomised trials of binocular treatment versus conventional patching or blurring in children aged 3 to 8 with unilateral amblyopia who had received only glasses beforehand and were followed for at least 8 weeks. It found one eligible trial and analysed the 68 children in it who met the age criterion (63 for the main acuity outcome). At 16 weeks, acuity improved by 0.21 logMAR with binocular treatment and 0.24 logMAR with patching, a difference of 0.03 logMAR (95% CI -0.10 to 0.04) that was neither statistically nor clinically significant. Adherence was higher with patching (90% of that group vs 47% of the iPad group achieved more than 75% compliance), and no permanent double vision was reported [9].

The authors were moderately confident that after 16 weeks the gain with binocular treatment is likely comparable to conventional patching, but said the small sample and lack of 52-week follow-up prevent robust conclusions about safety and lasting benefit. They called for more research with longer follow-up, user-reported outcomes, adherence data and information on recurrence after treatment stops [9]. Note that the review's strict criteria left out trials that did not fit (for example those enrolling previously treated children), which is one reason it found so few studies.

Regulation: what the FDA has authorized

On 20 October 2021 the US Food and Drug Administration granted a De Novo request for Luminopia One, classifying it as a Class II "digital therapy device for amblyopia" (21 CFR 886.5500, product code QQU). The request had been received on 1 March 2021 [14]. The FDA's indication reads, in summary:

• It is a software-only digital therapeutic used with commercially available head-mounted displays.
• It is indicated for improving visual acuity in people with amblyopia aged 4 to 7, associated with anisometropia and/or mild strabismus.
• Use follows treatment instructions (how often and how long) prescribed by a trained eye-care professional; it is for prescription use only, at home.
• It is intended for previously treated and untreated patients, but patients with more than 12 months of prior treatment (other than refractive correction) have not been studied.
• It is an adjunct to full-time glasses, which are also worn under the headset.

The trial of that product is the one described above [8]. This shows that regulators have accepted a randomised trial for one dichoptic therapy in one age group, on a specific headset. It does not extend to other software, other displays, older patients, or unsupervised use. ZyncView is not FDA-authorized, is not a medical device, and is not that product.

Adherence: the recurring problem

Whatever the treatment, it only works if it is used. In ATS18, only 22.2% of children in the game group completed more than 75% of the prescribed treatment [2]. In ATS19, only 13% did [6]. In the trial analysed by the Cochrane review, 47% of the iPad group reached that level, compared with 90% of the patching group [9]. The ATS19 authors explicitly said they could not tell whether a weak response came from a lack of effect or from a lack of play [6].

Game makers often argue that engaging design improves adherence. In the iPad-game trials above, adherence was still low. Enjoyment helps, but it needs to be measured and not assumed.

The practical lesson is that a treatment which is enjoyable and short enough to do every day may be worth more than one which is theoretically stronger but rarely used, and that adherence should be measured, not assumed. This is why ZyncView keeps sessions short and shows how many minutes you have played.

Adults and older children

The NEI notes that treatment is usually less effective in adults than in children [1]. The trials above enrolled children and teenagers: in teenagers a binocular game was not better than patching [6], and in previously treated children aged 7 to 12 it gave no extra benefit over glasses [3]. For adults, the laboratory evidence is more encouraging about plasticity [11][12], but in the sources reviewed here it is laboratory and early clinical data, not large trials of home games. If you are an adult with amblyopia, an eye examination can still be valuable, and your doctor can explain what is realistic.

What this means for ZyncView

• ZyncView is a different product from the ones tested. The PEDIG trials used specific iPad games; the FDA-authorized product uses a headset and video content. ZyncView is a browser-based set of wellness games and a reading exercise, used with red/blue glasses. It is not a medical device. Its effect has not been studied in a published trial, and we do not claim any health benefit.
• Games are not a replacement for eye care. The FDA-authorized product is itself an adjunct to glasses, and standard treatments remain the established options [1][14]. If you or your child may have a vision problem, see an eye care professional.
• Keep sessions short and regular. Adherence is the recurring weakness of trials [2][6], so ZyncView suggests 10 to 20 minutes a day, 5 days a week.
• Game scores are not health measurements. ZyncView records the level at which each session ends, scores and minutes. These numbers are game data, not a measurement of visual acuity.
• Stop and get advice if you have headaches, eye strain that does not settle, or double vision.

How to read a trial: a glossary

Visual acuity

How small a letter you can read on a chart. Measured separately for each eye.

logMAR and "lines"

A scale for acuity. 0.1 logMAR equals one line on the chart, which is 5 letters. Lower logMAR is better vision.

Amblyopic eye and fellow eye

The weaker eye and the stronger eye.

Suppression

The brain turning down the signal from one eye.

Dichoptic / binocular treatment

Showing different images to each eye at the same time so both eyes contribute to a task.

Anisometropia and strabismus

A large difference in focusing power between the eyes, and misaligned eyes. Common causes of amblyopia.

Stereopsis

Depth perception that depends on both eyes working together.

Randomised controlled trial

Participants are assigned to treatments by chance, so the groups are comparable.

Sham

A look-alike treatment with no expected effect, used as a comparison.

Non-inferiority

A trial designed to show a new treatment is not worse than the standard by more than a set margin (0.5 lines in ATS18). Failing to show it does not prove the new treatment is worse.

Confidence interval (CI)

A range that likely contains the true effect. If it includes zero difference, the result may be due to chance.

Masked examiner

The person measuring acuity does not know which treatment the child received.

Intention-to-treat

Analysing everyone as randomised, even if they did not follow the treatment.

Adherence (compliance)

How much of the prescribed treatment was actually done.

Questions to ask your eye doctor

• What is causing the amblyopia in my case (or my child's), and has that been corrected?
• What are the standard treatment options and what improvement is realistic?
• Would a dichoptic game be a useful addition for us, at our age and treatment history?
• How long and how often should sessions be, and how will we measure whether it helps?
• What should make us stop or call you (headaches, double vision, eye strain)?
• How does treatment interact with glasses, patching or drops already in use?

How this page was written

The ZyncView team wrote this page on 2026-09-26. We used the National Eye Institute's amblyopia page (last updated 26 November 2024) and the FDA's De Novo decision letter for Luminopia One. We also used the abstracts of the randomised trials, reviews and studies listed below, found by searching the Europe PMC index of biomedical literature. We favoured randomised trials and systematic reviews and quoted their own conclusions. This is a summary, not a systematic review: it does not cover every study, we may have missed relevant work, we only had access to abstracts, and it has not been reviewed by a named clinician. New research can change the picture; if you spot an error or a missing important study, please write to hello@zyncview.com and we will correct it.