start » program » presentations

Relation between motion perception and gaze direction. Evidence from VOG eye tracker


The relation between motion perception and eye movements is still controversial (Spering and Montagnini, 2011). Some findings concerning pursuit eye movements suggest that both common and distinct mechanisms guide perceptual and motor responses to visual motion information. However, these studies measure mainly eye movements velocity as compared to stimuli velocity. It is still unclear if different motion stimuli would elicit specific spatial patterns of gaze. This study aims to analyze how coherent motion perception is related to gaze direction by measuring eye movements’ spatial information.


The eye movements of eight healthy participants (mean age 19.9 years) have been recorded during a motion coherence discrimination test. The stimuli consists of 100 white dots (0.2°) moving at constant velocity on a black background. The area subtending the motion is a circular frame (7° of diameter). Each dot has a limited lifetime of 200 msec. The coherence level is held constant at 4 decibel (i.e. 36% of dots move coherently in one of eight possible direction, cardinal or oblique, while the rest moves in Brownian manner). Each subject performs eight tasks, one in each direction of motion. In order to measure the gaze pattern of exploration during the motion perception, the stimuli duration is set at 3 seconds. The subject is asked to discriminate the direction of coherent moving dots and his gaze position is recorded binocularly. For this purpose an eye tracker SMI 500 is employed. It is a non-invasive, image based system using a remote-controlled infrared eye camera with automatic eye and head tracker. The system maximum sampling frequency is 500Hz. In our experiment the 120Hz resolution is used.Horizontal and vertical components of gaze are represented in XY graph. Confidence ellipses of gaze are calculated as the area subtending the 95% of horizontal by vertical gaze position. The principal axe inclination is considered as indicative of the main gaze direction and it is compared to the direction of motion stimuli.


The gaze direction does not always correspond to stimuli motion direction even in those subjects who perceive correctly the motion. However, the evaluated confidence ellipses show that the principal axe is longer than the other axe suggesting that the stimulus is explored by specific directional eye movements and not by random gaze position. From these results it could be hypothesized that the space exploration of moving stimuli is specific to motion detection, confirming the important contribution of eye movement analysis in perception strategies investigation.