Abstract
Visual processing and visual perception are both thought to be tremendously depleted during saccadic eye movements. Using a projection system with sub-millisecond temporal resolution we rendered stimuli briefly visible during saccades and tested whether intra-saccadic localisation depends on retina-centered (retinotopic) or world-centered (spatiotopic) coordinates. Participants carried out large saccades in one of eight directions. After the cue to saccade, as indicated by the disappearance of the fixation markers, a stimulus was introduced that was invisible at fixation but visible during the saccade. This was achieved by drifting a Gabor within a stationary envelope in the direction of the eye movement and at a speed equal to the average peak velocity of the participant's own saccades (measured in a separate session). As a consequence of this manipulation, the grating would become stationary on the retina when the eyes were in mid-flight. The stimulus was extinguished at the end of the cued saccade. Participants either immediately made a second saccade to the perceived location of the stimulus or they used a mouse pointer to indicate the stimulus location. We found clear evidence that participants based their localisations unequivocally on spatiotopic coordinates—they indicated the actual location of the stimulus on the screen and were not influenced by the location the stimulus on the retina when it became visible during the saccade. Localisation performance was largely accurate to within one degree of visual angle, irrespective of whether participants made a saccade or used the mouse to indicate the location of the stimulus. The ability to access the world-centered location of purely intra-saccadic stimuli suggests that the position of the eye is accessible with high temporal and spatial fidelity, and readily used for perception and motor control.
Meeting abstract presented at VSS 2017