Automated visual acuity estimation by optokinetic nystagmus using a stepped sweep stimulus.

Abstract

Purpose: Measuring visual acuity (VA) can be challenging in adults with cognitive impairment and young children. We developed an automatic system for measuring VA using Optokinetic Nystagmus (OKN). Methods: VA-OKN and VA by ETDRS (VA-ETDRS) were measured monocularly in healthy participants (n=23, age 30±12). VA was classified as reduced (n=22, >0.2 logMAR) or not (n=24, ≤0.2 logMAR) in each eye. VA-OKN stimulus was an array of drifting (5 deg/sec) vanishing disks presented in descending/ascending size order (0.0 to 1.0 logMAR in 0.1 logMAR steps). The stimulus was stepped every 2 seconds, and 10 sweeps were shown per eye. Eye tracking data determined when OKN activity ceased (descending sweep) or began (ascending sweep) to give an automated sweep VA. Sweep traces were randomized and assessed by a reviewer blinded to VA-ETDRS. A final per sweep VA and VA-OKN was thereby determined. Results: A single randomly selected eye was used for analysis. VA deficit group: There was no significant difference between overall mean VA-OKN and VA-ETDRS (p>0.05, paired t-test) and the r2 statistic was 0.84. The 95% limits of agreement were 0.19 logMAR. No VA deficit group: There was a 0.24 logMAR bias between VA-OKN and VA-ETDRS and no correlation was found (r2 = 0.06). However, the overall sensitivity/specificity for classification was 100%. Conclusions: A robust correlation between VA-ETDRS and VA-OKN was found. The method correctly detected a VA deficit.