Abstract
The foveola is the retinal region free from capillaries where cones are most densely packed. Although it only covers a tiny portion of the visual field (~1 degrees), foveal representation is disproportionally large throughout the visual hierarchy. Yet little is known about foveal sensitivity, primarily because of the technical challenges inherent in studying vision at such a small scale. The continual presence of eye movements (often at the resolution limit of standard eye-trackers) and the difficulty in accurate gaze localization, have traditionally prevented testing of adjacent retinal locations in isolation and contributed to the assumption that foveal sensitivity is approximately uniform. Building upon recent advances in eye-tracking, gaze-contingent display control, and localization of the line of sight, here we mapped visual sensitivity at 13 locations within the foveola of healthy observers. As in standard perimetry tests, subjects were asked to detect a small probe (5 arcminutes) briefly displayed (50 ms) over a uniform background (10 cd/m2). Unlike standard perimetry, however, probes were confined to the central 1 degree of the visual field. Probe locations were updated in real-time to compensate for eye movements, which were continually monitored by means of a digital Dual Purkinje image eye-tracker. Subjects reported their responses via button presses and received auditory feedback. By systematically varying the contrast of the probes, we measured contrast sensitivity functions at each location. Our results show that sensitivity is not uniform within the foveola. While idiosyncratic differences occur, peak sensitivity is consistently shifted towards the temporal visual field and favors the horizontal meridian relative to the vertical one. Furthermore, sensitivity progressively deteriorates with increasing eccentricity from the locus of peak sensitivity.