Related Experiment Video
Updated: Jul 7, 2026
An Automated Method for Assessing Visual Acuity in Infants and Toddlers Using an Eye-Tracking System
Published on: March 17, 2023
Psychometric Performance of Children With Amblyopia During a Tablet-Based Adaptive Visual Acuity Assessment
Jan Skerswetat1,2, Zain Ahmed3, Nikhil Chaudhry3
1Department of Psychology, Northeastern University, Boston, Massachusetts, United States.
Purpose:
Visual acuity is a key metric for the diagnosis of amblyopia. Here, we introduce a self-administered near visual acuity (VA) test using the Angular Indication Measurement paradigm (AIM-VA) and evaluate it in children with and without amblyopia.
Methods:
Thirty-five children (age range, 5-16 years; 20 males) with (n = 22) and without (n = 13) amblyopia monocularly completed the AIM-VA twice within the same day and a Snellen-VA test. The AIM-VA presented three 4 × 4 charts of Landolt-C optotypes that adaptively varied in size and had randomized gap orientations. Orientation report errors as a function of optotype size were fit with a one-parameter (VA) model or a three-parameter (VA, slope, and response error) model. Bland-Altman analysis assessed test-retest repeatability and intertest agreement with Snellen-VA. Linear correlations and leave-one-out analyses explored relationships between psychometric parameters across groups and amblyopia subtypes.
Results:
Interocular acuity difference (IOD) was significantly larger in children with amblyopia than in controls for both AIM-VA and Snellen-VA. The one-parameter AIM-VA yielded repeatability comparable to previous reports, particularly in the amblyopic group, and showed no significant test-retest bias (P > 0.05). However, there was a systematic mean bias of lower logMAR estimates relative to Snellen-VA in the amblyopic group (P = 0.002), likely reflecting the difference in testing distance and task. Exploratory analysis revealed that shallower psychometric slopes were consistently associated with larger noise values across groups, with stronger parameter covariance in refractive than strabismic amblyopia.
Conclusions:
AIM-VA successfully detected amblyopic near-vision IOD. The one-parameter AIM-VA model showed no retest bias and comparable variance in this pediatric population. Exploratory psychometric parameters showed promise as additional indicators of functional deficits, with potential phenotypical differences between amblyopia subtypes warranting further investigation. AIM-VA is an adaptive, self-administered task that may be an effective tool for managing amblyopia.
