Abstract
Our eyes are always in motion. Even when we attempt to keep the gaze steady on an object of interest, microscopic eye movements, which include microsaccades and drifts, continually shift the stimulus on the retina. It has long been known that fixational eye movements differ significantly among observers. However, a systematic characterization of individual variability in the pattern of microscopic eye movements has not been reported in the literature. In this study, we used a DPI eye tracker, a high-resolution device which enables observation of both microsaccades and drifts, to study how the two types of fixational eye movements contribute to maintain fixation in human observers. Subjects (N = 14) were asked to maintain prolonged fixation in various conditions, including the presence/absence of a fixation cue (a 5′ dot) and the presence/absence of a textured background (a natural image). Fixation was most accurate with a fixation cue, irrespective of whether the background was textured or uniform. In this condition, the area containing the line of sight with 0.75 probability was on average 464 arcmin2 and varied across observers by more than a factor of 10, ranging from less than 100 to more than 1000 arcmin2. Among all the examined oculomotor parameters the best predictor of fixation was the speed of ocular drift; subjects with lower drift speed systematically exhibited more accurate fixation. Without a fixation cue, the 0.75 probability area of gaze increased by a factor of 2 to 4 depending on whether or not a textured background was present. In this condition, each observer maintained their own speed of drift, but drift covered a much larger area. Microsaccades were less frequent but substantially increased in amplitude and heavily contributed to the instability of fixation.
NIH R01 EY18363 and NSF BCS-0719849.