Fixational Eye Movements Act More Globally when Seeing Locally: A Motion-Integration Perspective

Yujie Wu,Mei Qiu,Dajun Xing
DOI: https://doi.org/10.1167/jov.20.11.717
2020-01-01
Journal of Vision
Abstract:Fixational eye movements are assumed to be linked to visual perception. However, an increasing number of studies show dissociations between human perception and eye movements. Here we examined whether involuntary fixational eye movements, including microsaccades and intersaccadic drift, behave consistently with our perception when human and monkeys watch visual stimuli that elicit either coherent motion perception or incoherent motion perception. We recorded eye movements of human and monkeys in a fixation task while showing them three different motion patterns: pattern 1 is a diamond frame (8 degree for size, 0.5 degree for line width) rotating around a fixation point; pattern 2 is the same rotating frame with its corners covered by four black static occluders (0.1 cd/m2 for luminance, 4 degree for each occluder size); pattern 3 is a rotating frame also same as pattern 1 with its corners covered by four static occluders whose luminance are the same as background luminance (37.9 cd/m2). All human participants perceived motion pattern 1 and 2 as coherent motion, and perceived pattern 3 as incoherent motion (four line segments moving in two orthogonal trajectories). However, under all three conditions, participants showed similar circular traces of fixational eye movements which were consistent with circular trajectories of global motion. Similar to human participants, 4 monkeys showed circular eye movement traces for all three motion patterns. Furthermore, we found both directions of microsaccades and drift velocity showed periodic circular patterns consistent with global motion under three conditions. Our findings suggest that both microsaccades and drift in fixational eye movements reflect motion information in a global manner without awareness, which leads to a dissociation between fixational eye movements and perception in a motion-integration perspective.
What problem does this paper attempt to address?