Binocular Rivalry Reveals a Potential Diagnostic Marker for Lazy Eye

Smith-Kettlewell’s Adrien Chopin and colleagues found that the rate at which perception flips between two competing images separated adults with amblyopia, or lazy eye, from adults with normal vision, and tracked with the clinical type of the condition.

A four-minute test of how perception shifts between the eyes distinguished adults with amblyopia, or lazy eye, from adults with normal vision, Smith-Kettlewell researchers report in a study published in Investigative Ophthalmology & Visual Science.

In the 11 adults with amblyopia the team tested, perception flipped between competing images at a sharply reduced rate compared with eight adults who had normal vision.

The measures also sorted participants by clinical subtype, separating anisometropic amblyopia from strabismic and mixed forms.

“What excited me most was seeing that just four minutes of tracking these perceptual switches allowed us to perfectly separate healthy vision from lazy eye, and even distinguish the underlying causes of the condition. There is almost no overlapping in the group distributions.”— Adrien Chopin, first author

What is lazy eye (amblyopia)?

Amblyopia is reduced vision in one eye that develops when the brain and that eye fail to work together properly in early childhood. The eye itself is usually healthy. The problem is in how the brain learns to use it.

It is common. A meta-analysis in the British Journal of Ophthalmology, pooling 60 earlier studies, put global prevalence at 1.44 percent and estimated 99.2 million people had amblyopia worldwide in 2019, rising to a projected 175.2 million by 2030.

The condition takes different forms. Anisometropic amblyopia arises when the two eyes differ sharply in focusing power, so one delivers a blurrier image than the other. Strabismic amblyopia arises from strabismus, a misalignment of the eyes, which sends the brain two images it cannot reconcile. Mixed amblyopia involves both.

What is binocular rivalry?

Binocular rivalry is what happens when each eye is shown a different image and the brain cannot merge them. Instead of blending the two, perception flips back and forth between them on its own, without the viewer choosing.

The phenomenon has been studied for well over a century. What had not been done, Chopin realized, was to revisit it in amblyopia with modern measures.

He had assumed the question was long settled. The classic studies, he said, were never reexamined using reversal rates, the timecourse across a trial, or the mixed states that appear when neither image fully wins.

He raised it with Aubrey Rossi, a student in the Levi Lab at UC Berkeley, who took up the question for her honors thesis.

How the binocular rivalry test for lazy eye works

Participants viewed competing striped patterns through a stereoscope, a device that presents a separate image to each eye, and reported continuously what they were seeing as their perception shifted. Sessions produced as much as 40 minutes of data per person.

Rather than track which eye was winning, the researchers counted the transitions, and separated them into two kinds.

A complete reversal is a clean switch from seeing one eye’s image to seeing the other’s.

An incomplete reversal passes through a mixed percept, a patchy in-between state in which fragments of both images are visible at once and neither has taken over, and comes back to the same initial eye’s image without landing on the other eye’s image.

Counting both, and timing them, gave a dynamic signature of how the two eyes were competing rather than a single snapshot of which one dominated.

Can a four-minute test detect lazy eye?

In this study, yes. Plotting the rate of complete reversals against the rate of incomplete reversals pulled the two groups apart with very little overlap, and about four minutes of data was enough to do it.

Participants with amblyopia also spent more of the time seeing the fellow eye’s image. Adding that measure, the proportion of time spent seeing one eye, separated the clinical subtypes, placing participants with anisometropic amblyopia in a different region from those with strabismic or mixed forms.

The direction of that split is itself informative. Participants with anisometropic amblyopia showed higher incomplete reversal rates than those with strabismic or mixed forms, which the authors read as a sign of higher binocular noise in anisometropia.

That matters because the subtype shapes what is happening in the brain, and until now no single behavioral measure has reliably told them apart.

How this compares with other binocular rivalry research at Smith-Kettlewell

Separate work at the institute has examined the same phenomenon from another angle.

Chuan Hou and Junxian Rao reported in Vision Research in 2025 that in anisometropic amblyopia the amblyopic eye rarely won out at small image sizes, that boosting the image’s size or contrast partially restored it, and that people with amblyopia spent more time in the mixed, in-between state than people with normal vision.

That study focused on how long each image stayed dominant. Chopin’s team focused on the switches between them, which proved the more sensitive discriminator, and extended the question beyond anisometropia to strabismic and mixed amblyopia.

Together the two describe the same altered competition between the eyes from different directions. Hou has also worked on amblyopia detection directly, using sweep visual evoked potential measures of vernier and grating acuity to identify the condition.

Catching lazy eye in children too young to describe what they see

Current amblyopia assessment can be lengthy, and it depends on what a patient reports seeing. That is a particular difficulty in young children, the group the researchers are ultimately aiming to screen.

The researchers said the reversal patterns could in principle be read from eye movements instead of button presses. That would make the approach workable for children too young to report reliably.

For researchers, the same measures open a window onto interocular suppression, the process by which the brain holds one eye’s input at bay, and onto neural noise in the visual system.

The study is small and was conducted in adults. Confirming the results will require larger groups and testing in children.

The separation the team saw, though, is unusually clean for a condition that has proven hard to capture with any single measure. “No other amblyopia biomarker comes close to that,” Chopin said.

A test that runs in four minutes, asks nothing of a child but that they watch, and reports not only whether amblyopia is present but which kind, would put screening within reach of far more of them. That is the case for pushing this line of work from 19 adults toward the children it was always meant to reach.

That is the goal of the team’s broader project, which is aimed at developing a rapid, objective vision screening tool for preschool children.