Purchase this article with an account.
Agostino Gibaldi, Martin Banks; Oculomotor and Perceptual Adaptation to Natural Scenes Statistics. Journal of Vision 2018;18(10):239. doi: https://doi.org/10.1167/18.10.239.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
As we explore the 3D environment with our eyes, binocular coordination is required to allow both eyes to land on the object of interest quickly and accurately. From an analysis of natural-scene statistics, we know that points in the upper visual field are likely to be farther than the current fixation point, that points in the lower field are likely to be nearer, and that points to the left and right of fixation are likely to be farther. We investigated whether binocular eye movements are biased toward landing on the most likely depths in different parts of the visual field. We measured where the eyes land in depth when making upward, downward, leftward, and rightward saccades without visual feedback (aka open loop). The results show that open-loop movements are divergent with upward, leftward, and rightward saccades, and convergent with downward saccades. In other words, the initial landing points for binocular, open-loop saccades are consistent with the statistics of natural scenes. We also investigated whether this oculomotor behavior is compatible with the positions of corresponding retinal points in the two eyes. The externalization of those points is the horopter, which is the locus of points in space where stereo depth is most accurate. We asked whether the landing points for binocular, open-loop saccades tend to be on the horopter. We found that they are, which means that when the eyes move to a new position, the observed convergence or divergence is biased toward placing the landing point near the horopter such that that point stimulates corresponding points. Our results show that the oculomotor system is adapted to the statistics of natural scenes.
Meeting abstract presented at VSS 2018
This PDF is available to Subscribers Only