Much research has focused on how perceptual, cognitive, and attentional processes modulate microsaccades, the small rapid gaze shifts that humans perform when at tempting to maintain steady gaze on a point. Yet the reasons why these fixational saccades occur in the first place have remained unclear. Long-standing theories have argued for either spatial (i.e., gaze centering) or temporal mechanisms (i.e., a periodical release process). However, this debate has never been resolved, primarily because of uncertainty in deter mining where the observer looks. Whereas modern eye-trackers enable detection of small eye movements, accurate localization of the line of sight remains challenging. Here, rather than indirectly inferring gaze position from oculomotor activity, we used a gaze-contingent procedure to directly estimate the perceived center of the visual field, a method that has been previously shown to effectively reduce uncertainty. Our results from subjects of both sexes show that the generation of fixational saccades depends on the interaction of spatial and temporal factors. Fixational saccades are remarkably accurate in correcting for fixation errors, even when gaze is minimally displaced. However, fixational saccades also occur when gaze is centered, but their latency increases as the fixation error decreases. These results suggest that fixational saccades serve an important corrective function when needed, but they can only be avoided for a limited period of time when fixation is already accurate. Significance statement It is unclear why humans perform small rapid gaze shifts (microsaccades) while attempting to maintain steady gaze on a point. Building upon recent advances in eye-tracking and gaze-contingent display control, here we show that normal fixation requires the interplay between two sophisticated processes: the generation of very accurate targeted microsaccades and the selective suppression or delaying of unnecessary ones. These results advance knowledge on the visuomotor strategy by which humans maintain steady gaze and raise questions about how the interplay between these two processes is altered in conditions of impaired fixation.
Wang et al. (Wed,) studied this question.