Purpose: Recent advances in retinal imaging allow researchers to simultaneously stimulate and image the retina with a resolution on the order of one cone diameter. We have exploited this capability to examine how fixational eye movements influence motion discrimination judgments. Specifically, we examined up vs down motion

discrimination under referenced and unreferenced conditions, and we imaged the retina with an Adaptive Optics Scanning Laser Ophthalmoscope (AOSLO) so that the exact retinal position of the stimulus was known on each trial.

Methods: A single-interval, forced-choice psychophysical procedure was used to measure up/down discrimination thresholds for displacement of small, foveally viewed bar targets. Targets were 3 by 19 arcmin dark bars on a bright background (referenced motion condition) or 5 by 40 arcmin bright bars in an otherwise dark environment (unreferenced motion condition). AOSLO video sequences were recorded for off-line registration of retinal images. Psychometric functions and threshold determinations were then made for spatiotopic (bar motion in image) and retinotopic (bar motion on retina) frames of reference.

Results: Referenced motion thresholds were always better when expressed in a spatiotopic frame of reference, indicating that eye movements were compensated for in the judgments. Unreferenced motion showed the opposite relationship, with retinotopic judgments being generally more accurate than spatiotopic judgments, but subjects were still able to make some compensation for eye motion.

Conclusions: Subjects make precise judgments of displacement in spite of relatively large retinal image motions due to fixation jitter, as long as there is a frame of reference. Even without a frame of reference, subjects show some compensation for their eye movements indicating that some source of extraretinal information about fixational jitter is available.