Susana T. L. Chung¹ (), Sunwoo Kwon¹, Dennis M. Levi¹; ¹University of California, Berkeley

Our eyes are always moving even when we attempt to maintain steady fixation on a visual object. The retinal image motion associated with normal fixational eye movements (FEMs) has been shown to be important, yet not optimal, for seeing fine spatial details (Agaoglu et al., 2018). Because FEMs are more unsteady and are associated with larger amplitudes and higher velocities in eyes with amblyopia, we asked whether the naturally occurring FEMs are optimal for seeing fine spatial details in amblyopic eyes. Using a tracking scanning laser ophthalmoscope that is capable of delivering visual stimulus at precise retinal locations, we asked five observers with amblyopia to discriminate the orientation (±45°) of a sinusoidal grating presented at the fovea, for a spatial frequency 2.5× lower than the equivalent grating acuity of the eye. On each trial, the grating was presented at one of five stabilization gains (ratios of stimulus to eye velocity): -1, 0, 0.5, 1, 2. The contrast of the grating was adjusted such that observer’s performance was ~80% correct under the natural viewing condition (gain=0), for the fellow and amblyopic eyes separately. Eye and stimulus positions were extracted from recorded videos using a cross-correlation method. For all observers, task performance was the highest at gain=0 (no stabilization) and dropped at other gains (e.g. performance accuracies for gains of 0, -1 and 2 averaged 78.9%, 49.8% and 49.8%, respectively, in the fellow eyes, and 82.1%, 55.1% and 55.7%, respectively, in the amblyopic eyes). Contrary to our previous finding that in the normal fovea, the best performance was obtained at a gain of 0.43, here we found that amblyopic observers’ performance was the best with no stabilization, in both the fellow and amblyopic eyes, suggesting that fine spatial vision is optimally adapted to naturally occurring (abnormal) FEMs in the presence of amblyopia.

Acknowledgements: Supported by NIH Grants EY030609 (DML) and EY 030253 (STLC) and a postdoc fellowship from CIVO (SK).