VR for Amblyopia: What Dichoptic and Binocular Treatment Trials Show

Updated 2026-10-21

Written by the Optics Trainer team from published research, with every source linked below.

Virtual reality and other dichoptic treatments for amblyopia have moved from research into prescription devices in a few years, and parents now ask about them by name. This guide summarises what the randomised trials, the 2022 Cochrane review and the FDA records actually show: which devices are authorised and for whom, how they compared with glasses and patching, where the evidence is weak, and what is still being tested. It is written for clinicians deciding how to discuss these options with families.

Why binocular treatment for amblyopia

Glasses, patching and atropine remain the standard treatments for amblyopia, as the American Academy of Ophthalmology describes. Patching dose trials by the Pediatric Eye Disease Investigator Group (PEDIG) found that in children under 7, 2 hours a day worked as well as 6 hours for moderate amblyopia, and 6 hours as well as full-time patching for severe amblyopia, at 4 months.

Patching treats one eye at a time. Reviews by Eileen Birch and by Robert Hess and Ben Thompson describe a growing view that abnormal binocular interaction and suppression of the amblyopic eye are central to the condition. Birch also notes that many treated children are left with residual deficits or recurrence. Binocular and dichoptic treatments follow from that view: each eye sees a different image, and the image to the fellow eye is degraded or reduced in contrast so that the amblyopic eye has to contribute.

Early binocular game trials

PEDIG's 2016 trial randomised 385 children aged 5 to 12 to a binocular iPad game for 1 hour a day or patching for 2 hours a day for 16 weeks. Acuity improved 1.05 lines with the game and 1.35 lines with patching. The result did not show the game was as good as patching, and a post hoc analysis suggested it was not. Adherence was the main problem: only 22% of children with game logs completed more than 75% of the prescribed play.

A 2019 PEDIG trial of the Dig Rush game in 138 children aged 7 to 12, almost all previously treated, found no acuity or stereoacuity benefit over glasses alone after 8 weeks. In BRAVO, a placebo-controlled trial of 115 older children, teenagers and adults (mean age 21.5), a dichoptic falling-blocks game did no better than a placebo game over 6 weeks.

Dichoptic VR: the Luminopia trial

Luminopia presents children's TV shows in a VR headset, with complementary parts of the image and reduced contrast to the fellow eye. Its pivotal randomised trial (Xiao and colleagues, 2022) enrolled 105 children aged 4 to 7 at 21 US sites. The treatment group used the headset at home for 1 hour a day, 6 days a week, while wearing glasses full time. The control group wore glasses alone.

At 12 weeks, amblyopic-eye acuity improved 1.8 lines with treatment and 0.8 lines with glasses alone, a difference of 1.0 line. The trial stopped early for success at a planned interim analysis. Two points matter when explaining it to families: the comparison was with glasses alone, not patching, and the primary outcome was at 12 weeks.

Eye-tracking dichoptic treatment: the CureSight trial

CureSight (NovaSight) is screen-based rather than VR: it combines dichoptic video, viewed through red/blue glasses, with eye tracking. Its randomised trial (Wygnanski-Jaffe and colleagues, 2023) compared it (90 minutes a day, 5 days a week) with patching for 2 hours a day in 103 children aged 4 to under 9. At 16 weeks, acuity improved 0.28 logMAR with CureSight and 0.23 logMAR with patching, which met the trial's noninferiority margin. Adherence was 91% with CureSight and 83% with patching. Several authors are affiliated with NovaSight.

FDA status, ages and indications

Both devices are prescription-only, used at home, and indicated as an adjunct to full-time refractive correction for amblyopia associated with anisometropia and/or mild strabismus. They are FDA-authorised through De Novo and 510(k) clearance, not PMA approval:

Luminopia One: De Novo DEN210005, granted October 2021 for ages 4 to 7 (children with more than 12 months of prior treatment other than glasses were not studied). It created the device type "digital therapy device for amblyopia" (product code QQU, 21 CFR 886.5500, Class II).
Luminopia: 510(k) K243819, April 2025, widened the age range to 4 to under 13. The expansion relied on real-world registry data from 290 patients, not a new randomised trial. Two earlier clearances (K221659 and K233720) added headsets.
CureSight-CS100: 510(k) K221375, September 2022, for ages 4 to under 9.

What the systematic evidence says

The 2022 Cochrane review of binocular treatment versus patching in children aged 3 to 8 found only one eligible randomised trial, the PEDIG iPad study; the review's searches (to November 2020) predate the Luminopia and CureSight trials. In that subset of children it found moderate-certainty evidence of similar acuity gains at 16 weeks. It found no 52-week data, and concluded that firm conclusions about lasting effect and safety are not yet possible.

Evidence on durability is limited. A registry funded by Luminopia reported that gains held about 6 months after treatment ended in 60 children, with 3% recurrence (a loss of more than 2 lines), but the data are retrospective. In the registry data behind the 2025 age expansion, about 24% of patients showed no change and about 10% lost one or more lines.

Adherence and side effects

Adherence varies with the product and the population. It was poor with the iPad game in the 2016 PEDIG trial and better with CureSight in its trial. These are different devices, so results do not transfer from one to another.

In Luminopia's pivotal trial, treatment-related adverse events occurred in 25% of the treatment group and 13.6% of controls. Headaches were the most common (14.3% versus 1.7%), and all were mild and resolved. The FDA summary also lists diplopia, eye strain and nausea as risks.

What is still being tested

The head-to-head question is open for Luminopia. As of February 2026, PEDIG was recruiting about 238 children aged 4 to 7 for a randomised trial of Luminopia (1 hour a day, 6 days a week) against patching (2 hours a day, 7 days a week). The primary outcome is acuity at 26 weeks, and primary completion is estimated for July 2027.

When talking with families, the evidence supports a few clear points: these are prescription treatments for specific ages and diagnoses; Luminopia was tested against glasses alone, and CureSight against patching; binocular games tested in previously treated 7- to 12-year-olds and in older patients did no better than glasses alone or a placebo game; and long-term data are still limited.

Where Optics Trainer VR fits

Optics Trainer VR runs vision training activities on a Meta Quest headset (Quest 2 or newer), controlled from a tablet or clinic browser. Clinicians can see the patient's view and adjust difficulty, target eye, vergence, yoke prism, torsion and per-eye contrast where an activity supports them, across vergence, stereo acuity, motor fusion and training games.

It is not one of the FDA-authorised amblyopia devices described above, it has not been studied in these trials, and none of their results apply to it. It is not a treatment for amblyopia and does not replace prescribed treatment.

Frequently asked questions

Is VR treatment for lazy eye FDA approved?

Not "approved" in the PMA sense. As of October 2026, the FDA lists two dichoptic devices for amblyopia with anisometropia and/or mild strabismus, authorised through De Novo and 510(k) clearance: Luminopia, a VR headset (ages 4 to under 13 since April 2025), and CureSight, a screen with eye tracking and red/blue glasses rather than VR (ages 4 to under 9). Both are prescription-only and used at home with full-time glasses.

Is VR treatment as good as patching?

It depends on the device. Luminopia, a VR headset, was compared with glasses alone; a PEDIG trial against patching is under way, with primary results estimated for mid-2027. CureSight, a screen-based dichoptic device rather than VR, was noninferior to 2 hours a day of patching at 16 weeks. A binocular iPad game was not shown to be as good as patching.

Does VR amblyopia treatment work for adults?

There is not enough evidence to say. The main placebo-controlled trial in older patients, BRAVO (mean age 21.5), tested one dichoptic game on an iPod Touch, not a VR headset, and it did no better than a placebo game. The FDA-authorised devices are indicated only for children, up to under 13.

How long do gains from dichoptic treatment last?

Long-term data are limited. The 2022 Cochrane review found no 52-week data, and the main durability report for Luminopia so far is a retrospective, industry-funded registry showing gains held about 6 months after treatment.

What are the side effects of VR amblyopia treatment?

In Luminopia's pivotal trial, mild headaches were more common than with glasses alone. The FDA summary also lists diplopia, eye strain and nausea as risks.