Abstract
Visual tracking movements consist of two very different components: saccades and smooth pursuit. Under natural conditions, we frequently initiate a saccade to a peripheral object when it begins to move. This situation, where the direction of the saccade and subsequent pursuit are different, has not yet been systematically investigated under laboratory conditions. A single Gaussian blob appeared 5 deg eccentrically relative to a central fixation spot, either at the same vertical position or diagonally at an angle of 30 deg upward or downward. After the observer fixated the target for 500 ms, it made a step back towards the center of the display, and then moved at 20 deg/s in the direction opposite of the initial position, randomly in one of three directions, either horizontal or diagonal at 30 deg degrees upward or downward. Eye position was measured with an EyeLink 1000 system. Observers typically made initial saccades towards the center and then followed the target by smooth pursuit. The exact saccade landing position depended heavily on the subsequent pursuit direction. Saccades were displaced in the direction of the pursuit target. This effect increased with saccadic latency. We found three benefits when the direction of initial saccades and subsequent pursuit were identical: post saccadic position errors between the saccadic landing position and the moving target were smallest, post saccadic pursuit directional precision was highest and post saccadic pursuit speed after initial saccades were higher by 1 to 3 deg/s, This effect was strongest for the purely horizontal trials. We conclude that saccades and pursuit interact dynamically to optimize the following of sudden peripheral motion onsets.
Meeting abstract presented at VSS 2016