Ângela Gomes Tomaz¹, Sunwoo Kwon¹, Dennis M. Levi¹, Wolf M. Harmening², Bilge Sayim³; ¹Herbert Wertheim School of Optometry & Vision Science, University of California, Berkeley, Berkeley, United States of America, ²Adaptive Optics and Vision Laboratory, Rheinische Friedrich-Wilhelms-Universität Bonn, Bonn, Germany, ³Cognitive and Affective Sciences Lab (SCALab), UMR CNRS 9193, University of Lille, Lille, France

Persons with amblyopia, a neurodevelopmental disorder of spatial vision characterized by reduced visual acuity (VA), perceive stimuli distorted (e.g., Barrett et al., 2002, Pugh, 1958) and have stronger foveal crowding (i.e., worse performance with flanked targets) than neurotypical controls. Here, we explored the appearance of both crowded (with black bars on all sides) and isolated high contrast letters in observers with amblyopia and controls. Stimuli were presented in the fovea for 500ms with stimulus size set at 1.0 and 1.5 x the observer’s VA threshold (measured with isolated targets). Observers viewed the target monocularly with their dominant/fellow (DE/FE) or non-dominant/amblyopic (NDE/AE) eye and replicated its appearance on a 9x9 square-grid interface with binocular viewing. Each square in the interface could be turned on and off with mouse clicks and their grey levels adjusted with the mouse wheel. Results showed that response-target structural similarity (RTSS) was lower for 1.0 than 1.5 x threshold sized stimuli, for flanked than isolated targets and for observers with amblyopia than controls. No differences were found between DE/FE and NDE/AE viewing conditions in amblyopic observers. However, their RTSS in both eye conditions were lower than controls’ NDE condition. Surprisingly, the controls’ RTSS was lower with DE than with NDE viewing. There was no interaction between groups (Amblyopic, Controls) and crowding conditions (flanked, isolated). When compared to the target letter, responses revealed truncation and extension of elements, fusion of separate elements (e.g., connecting the flankers) and shape distortions (e.g., depicting straight lines as curves) for both groups in all conditions. Taken together, our results reveal characteristics of target appearance in both groups showing stronger visual distortions in persons with amblyopia than controls, but no crowding differences between groups. These results highlight important appearance differences and similarities between normal and amblyopic visual perception when controlling for VA differences.