Dichoptic therapy, and what the trials actually show
Dichoptix is a dichoptic training program used alongside the care you provide. This page sets out the mechanism, the randomised evidence with its numbers, where the approach fits in a treatment plan, and how consent-based monitoring works. It is written for clinicians, so it states limits as plainly as findings.
Equivalent acuity gains to patching, on roughly half the prescribed hours and with measurably higher adherence.
- 91%
- Treatment adherence in the dichoptic video arm, against 83% for patching in the same trial.
- 120 / 224
- Prescribed hours across 16 weeks, dichoptic against patching, for a result that was not inferior.
- 4.2
- Hours of patching actually delivered where 6 hours a day were prescribed, under objective monitoring.
Wygnanski-Jaffe et al., Ophthalmology, 2023
Wygnanski-Jaffe et al., Ophthalmology, 2023
Stewart et al., BMJ, 2007
The trial was a non-inferiority design: dichoptic treatment was not shown to be superior to patching on visual acuity, it was shown not to be worse. The demonstrated advantage lies in adherence and in the hours required.
Treating the binocular layer, not only acuity
Amblyopia carries a loss of acuity and a disorder of binocular function: suppression, weak fusion, reduced stereopsis. Monocular treatment addresses the first. Dichoptic presentation addresses both by giving each eye a separate but complementary stimulus that only combines when both contribute.
- 01
- Amblyopic eye
- 02
- Fellow eyeContrast balance
- 03
- Combined percept
Contrast is set separately per eye. Lowering the fellow eye's contrast lets the amblyopic eye's signal enter perception; as balance improves, the difference is reduced towards equality.
Suppression is the target
The task is built so it cannot be completed with one eye. The weaker eye's input has to be used, which is what patching cannot ask for by design.
Contrast is the control
The fellow eye is attenuated only as much as fusion requires, and the attenuation is reduced as balance improves. This is the same lever the binocular rivalry literature describes.
Optical correction comes first
Dichoptic work is done wearing the correction and after refractive adaptation, never instead of it. Correction alone resolves a meaningful share of anisometropic amblyopia.
Three randomised trials, and what separates them
Two pivotal trials support dichoptic treatment; a third, built on a single game, fell short. Reading them together is more informative than reading any one of them.
| Trial | Design | Population | Comparator | Primary result |
|---|---|---|---|---|
| Luminopia (Xiao 2022) | Phase 3 RCT, 21 sites | 105 children, 4 to 7 years | Full-time glasses alone | Amblyopic eye 1.8 lines vs 0.8 lines at 12 weeks; difference 1.0 line, stopped early for success. |
| CureSight (Wygnanski-Jaffe 2023) | Multicentre RCT, non-inferiority | 103 children, 4 to under 9 years | Patching, 2 h/day | 0.28 vs 0.23 logMAR, non-inferior. Stereoacuity improved 0.40 log arcsec in both arms. |
| Binocular iPad game (Holmes 2016) | Multicentre RCT, non-inferiority | 385 children, 5 to under 13 years | Patching, 2 h/day | 1.05 vs 1.35 lines. Non-inferiority indeterminate; post hoc favoured patching. |
What separates them is adherence
Every trial above delivered the same dichoptic principle. What differed was whether the content could be sustained, and adherence tracks the results closely.
- Single binocular game46%
median, 22% passed three quarters (Holmes 2016)
- Patching, 12 h/day prescribed52%
6.2 h received (Stewart 2007)
- Patching, 6 h/day prescribed70%
4.2 h received (Stewart 2007)
- Patching arm, CureSight trial83%
2 h/day prescribed (Wygnanski-Jaffe 2023)
- Dichoptic video treatment91%
90 min/day, 5 days/week (Wygnanski-Jaffe 2023)
The two patching percentages are derived from the reported mean hours received against hours prescribed.
Where the approach fits
The frame below is how we describe the program's scope. Indication, timing and the treatment plan remain yours.
Reasonable candidates
- Anisometropic, small-angle strabismic or mixed-mechanism amblyopia
- Refractive correction prescribed and worn, with refractive adaptation allowed to run its course
- A child able to cooperate with a screen task for short daily sessions
- A household able to sustain a daily routine over months
- Adherence with patching or atropine has broken down, or the binocular deficit is the priority
Assess before recommending
- Deprivation causes not yet addressed, such as cataract or significant ptosis
- Large-angle manifest strabismus, where alignment is the first question
- Photosensitive epilepsy or any contraindication to screen-based tasks
- Children too young to cooperate reliably with the task
- Nystagmus or ocular pathology requiring its own management first
The one precondition
Cycloplegic refraction and full-time correction come first. In previously untreated anisometropic amblyopia, correction alone improved acuity by two lines or more in 77% of children and resolved it in 27%, with gains continuing up to 30 weeks (Cotter 2006). Adding therapy before that has run its course attributes to the therapy what the glasses were going to do.
What to measure, and what each measure answers
Acuity remains the primary outcome, but a binocular treatment with only a monocular outcome measure reports on part of its own target.
| Measure | What it answers | Suggested cadence |
|---|---|---|
| Visual acuity, linear optotypes | Depth of amblyopia and the interocular difference, which matters more than the absolute value. | Every visit, same chart and distance |
| Stereoacuity | Binocular function, the outcome a dichoptic approach is aimed at. Can move independently of acuity. | Baseline, then every 2 to 3 months |
| Suppression and binocular balance | Whether the weaker eye's signal is entering perception at all. | Baseline and follow-up |
| Contrast sensitivity | Useful where acuity has plateaued while functional complaints persist. | As indicated, not routine |
| Delivered dose, from session logs | At a plateau, separates non-response from insufficient treatment. Self-report cannot. | Continuous, visible in the panel |
Changing optotype system between visits is the most common source of apparent change that is not real. Record the test, the distance and the correction status beside the value.
Consent-based access, logged end to end
You see a patient's metrics only while the link is active and consent stands. Nothing is visible before the handshake completes, and every transition is written to an append-only audit trail.
- 01
You hold a connect code
Each supervisor account carries its own code. Nothing is shared until a patient uses it.
- 02
The patient redeems it
Entering the code creates a pending request and records the patient's explicit consent to share therapy metrics. The screen states plainly what is being shared.
- 03
You accept or decline
The link becomes active only on your acceptance. Metrics become visible at that point and not before.
- 04
Either side can end it
The patient can withdraw consent at any time without ending the link, or remove the link outright. Visibility stops immediately in both cases.
What you see
Session frequency and duration, in-game accuracy and reaction time, difficulty settings, and the trend across weeks. What you do not see is anything from before the link was accepted, or anything after consent is withdrawn.
KVKK
Therapy metrics are personal health data. Consent is explicit, revocable and auditable by design, and the record is held in the EU.
How to get a clinician account
Clinician accounts are not self-serve. We enable them individually, so that a supervisor account always belongs to a verified eye-care professional. Write to us with your name, your practice and your registration details, and we will set it up and send your connect code.
info@dichoptix.comSelf-registration through the ordinary sign-up form creates a patient account, not a supervisor one.
The clinical shelf of the Learning Center
Longer treatments of the same material, each with its citations.
The evidence for dichoptic treatment: where we stand today
A summary for clinicians: the pivotal randomised trials, the head-to-head comparison with patching, and the adherence data. The decisive finding is that the same principle ran at 46% adherence as a single game and 91% as rich video content.
ReadPatient selection and follow-up in dichoptic treatment: a practical frame
For clinicians: which patients a binocular approach suits, what to assess at baseline, which measures to track in follow-up, and when to revisit the plan.
ReadOutcome measures in amblyopia: what to track, and when
Visual acuity is not the whole of follow-up. Which measure answers which question, the errors that destroy comparability, and what to track in a dichoptic program.
Read
Dichoptix is a therapy-support tool used alongside professional eye care. It is not a medical device, it is not CE marked or FDA cleared, and no efficacy claim is made for this program. The trials cited above evaluated other products and are reported here as the evidence base for the approach, not as results for Dichoptix. Diagnosis, the treatment plan and all clinical decisions remain yours.