Abstract
Even during fixation, our eyes are not perfectly steady but instead make small-scale eye movements (microsaccades) that have been proposed to be task related. Large intentional saccades to a peripheral target that is crowded (surrounded by flankers) have been shown to reduce the harmful effects of crowding. However, the impact of fixational eye movements (FEM) on crowding remains unknown. Here, we investigated fixational eye movements when performing a crowding task. A tracking scanning laser ophthalmoscope (TSLO) with a 3° raster and an 840 nm imaging beam was used to image the right eye of each observer (4 males and 1 female with normal ocular health) during a psychophysical task. Stimuli were presented on an LCD display (screen size of 62.5 cm, 1920 X1080 pixels) optically overlaid with the imaging raster. The stimuli consisted of Sloan numbers (0-9) presented for 100 ms, either unflanked or surrounded by Sloan numbers at one of 4 nominal spacings. This was repeated for eccentricities 1 through 5 degrees, with sizes scaled to provide 75% performance in the unflanked condition at each eccentricity. Strip-based cross-correlation was used to retrieve eye motion from the retinal videos. 480 retinal videos (2 minutes each) were analyzed, from which 48,533 microsaccades were detected. Strong suppression of microsaccades was observed 150 ms following stimulus onset. Interestingly, in the critical period between 150 and 250 ms following stimulus onset, average microsaccade amplitudes systematically decreased as eccentricity increased (0.014°/degree of eccentricity). When flankers were absent or at the largest spacing, directional congruency of microsaccades in relation to the target in the 200-400 ms time bin was predictive of behavioral performance, but this effect was abolished when flankers were closer in. We found that in a peripheral flanked identification task, performance-related microsaccadic activity was less pronounced for crowded stimuli.