This is a multi-center, randomized controlled trial to compare the effectiveness of AR training with patching for the treatment of unilateral amblyopia.

Specific Aim 1 (Primary): To compare the improvement of visual acuity in the amblyopic eye between AR training and patching for the treatment of unilateral amblyopia.

Specific Aim 2 (Secondary): To compare the improvement of visual functions between AR training and patching for the treatment of unilateral amblyopia.

Poor compliance, limited improvement of visual functions, and regression after recovery of visual acuity have been observed in the management of amblyopia using conventional patching. Recently, dichoptic/binocular digital therapy has been developed, but no widely accepted binocular treatments with superiority available for children and adults with amblyopia (Pineles et al., 2020; Oscar et al., 2023). Here, we designed an innovative binocular therapy using augmented reality (AR) training, based on neural deficits in amblyopia, to achieve better outcomes.

Selective deficits were found in the parvocellular pathway (P pathway) compared to the magnocellular pathway (M pathway) in the monocular processing of visual information in the amblyopic eye (AE) (Wen et al., 2021). In addition to monocular deficits, imbalanced binocular suppression may also play important roles in the visual deficits of amblyopia as suggested by clinical evidence (DeSantis, 2014; Von Noorden, 1996) and psychophysical studies (Baker et al., 2008; Holopigian et al., 1988; Li et al., 2011; Zhou et al., 2013). Based on the neural deficits in unilateral amblyopia, we first apply the push-pull approach (Xu, He & Ooi, 2010; Ooi et al., 2013), which was aimed to reduce sensory eye dominance in previous literatures, into the rebalance of functions of M and P pathways in the AE and the rebalance of binocular interaction, to improve the high spatial detail perception of the AE in daily life under binocular viewing condition, as well as binocular functions.

Using AR technique combined with dichoptic device, we present differentially-processed images to each eye of the patients in real time, allowing them to interact with the surrounding environment during the visual training. Using a Butterworth filter with the cutoff at 2 cycle pre degree, the images captured in real time are divided into information with high and low spatial frequencies (SFs) corresponding to the P and M pathways, respectively. For the AE, original low SF phase of captured images is scrambled into random noise with the refresh rate of the display, while the original information with high SF is retained completely. As a result, the function of the P pathway is pulled while the function of the M pathway is pushed, actively encouraging the interaction with the surrounding environment through high SF information. For the fellow eye (FE), original high SF phase of captured images is scrambled into random noise with increased contrast and reduced temporal frequency, while the contrast of the original high SF information is reduced. As a result, in addition to the push-pull in monocular P&M pathways, the function of the P pathway in the FE is pulled and while the function of the P pathway in the AE is pushed, actively improving the rebalance of binocular inhibition.

The proposed trial will be conducted in 4 different study sites in China. For the AR training group, patients need to perform AR training for 2 hours per day at home. For the patching group, patients need to patch the FE for 2 hours per day at home.

Inclusion Criteria:

1. Aged 5-55 years (including 5 years and 55 years);
2. Best-corrected visual acuity worse than 20/30 but no worse than 20/200 in the amblyopic eye, interocular difference of 2 or more lines, with the better eye within the normal range;
3. Patients have applied optical refractive correction for more than 3 months;
4. Ability attend visits and complete the treatment;
5. Normal binocular alignment including strabismic amblyopia with orthotropia after surgical correction, or intermittent exotropia within a range of -15 to 0 prism diopters measured by the prism cover test.

Exclusion Criteria:

1. Organic eye diseases preventing the establishment of good vision (e.g. ptosis, media opacity, nystagmus, paracentral fixation, acute inflammation like keratitis, optic nerve diseases like glaucoma, retinal diseases);
2. Lesions of the brain preventing the establishment of good vision (e.g. cortical visual impairment);
3. Implantable electronic device;
4. A history of ocular surgery (except strabismus surgery) affecting vision (e.g. retinal detachment repair);
5. A history of ocular trauma affecting vision;
6. Receiving amblyopia therapy (except wearing glasses) within 2 weeks prior to presentation;
7. History of epilepsy or mental illness, or cognitive defects;
8. Currently taking medications or needing to take medications during the study period that may affect vision;
9. Inability to comply with the treatments or follow-up visits required;
10. Participation in clinical trials on drugs within 3 months prior to presentation, or clinical trials on other medical devices within 30 days prior to presentation.

We concerns about patient privacy issues and it's better to protect the publication potential.