Purchase this article with an account.
C. Eccelpoel, K. Verfaillie; Incidental vs. deliberate coding of object orientation across eye movements. Journal of Vision 2001;1(3):221. doi: https://doi.org/10.1167/1.3.221.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
When an object is viewed across a saccade, intrasaccadic orientation changes are easier to detect than intrasaccadic position changes. This could mean that transsaccadic integration subserves object identification and that object orientation, unlike object position, is an integral part of the object representations that are matched to achieve identification. To test this hypothesis, we started from the observation that object semantics interfere with object matching whenever matching requires a shape judgement. According to Boucart and Humphreys (1997) this is the case because shape is represented in the object lexicon which interfaces object identity and semantics. Inspired by this account, we hypothesized that object semantics should modulate the detection of intrassaccadic changes in object orientation, but not in object position. To test this, viewers saccaded from one object to another and had to detect intrasaccadic changes in object position or orientation. The objects were categorically related or unrelated. Semantics affected processing in the orientation change detection task but not in the position change detection task. This supports our hypothesis that object orientation is part of the intermediate representations involved in transsaccadic object identification. To investigate whether orientation is automatically coded or only plays a role when viewers are explicitly monitoring orientation changes, we conducted a study in which orientation was irrelevant to performance and sensitivity to intrasaccadic orientation changes was measured implicitly. Specifically, viewers inspected arrays of tinted objects to decide which of the objects had the warmest/coldest colour. During inspection, intrasaccadic orientation changes were executed and subsequent object fixation times were analyzed as a function of orientation continuity and semantic relations between objects. The results are discussed in terms of their implications for viewpoint-dependent accounts of object recognition across saccades.
This PDF is available to Subscribers Only