Purchase this article with an account.
Jun-Yun Zhang, Lei Liu, Cong Yu; Evidence for misplaced target information with letter crowding. Journal of Vision 2008;8(6):441. doi: 10.1167/8.6.441.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Purpose: Recent debates on mechanisms of visual crowding are focused on whether crowding results from lower-level target-flanker feature binding, or higher-level insufficient resolution of attention. We took a novel route to understand crowding by investigating whether and where the correct target information could be traced when wrong responses are made. Methods: A letter target with flankers was presented in the right visual hemi-field (10° eccentricity). Subjects reported either the central target letter (partial report) or all three letters from left to right (full report). Results: 1) With the target size set substantially sub-threshold, full-report data suggest: a. The correct report rate of the target at the center location was comparable to that with partial report; b. Report of the center target at the right flanker location was substantial; c. So was the left flanker being reported at the center location. 2) Halving the left flanker contrast increased the center report rate of the left flanker, which, however, unchanged the center report rate of the center target. 3) Restricting stimulus misplacement by assigning half letters as target and half letters as flankers, which the subjects knew, reduced crowding by 30–40% in partial report. 4) Under the same conditions of 3), crowding became much weaker when flankers were pixel-scrambled (no more features), but was stronger when flankers were stroke-scrambled (intact features but broken Gestalt). Conclusions: 1) A significant portion of target information is not lost during crowding, but is positionally confused with flankers, presumably in the visual memory. 2) Flanker misplacement (Strasburger, 2002) had little effect on pure target identification. Center target identification under crowding may be earlier than flanker, and probably also center target, misplacement. 3) Top-down influences which control target misplacement improve correct report rate. 3) Lower-level stimulus configurations affect stimulus identification under crowding, consistent with the feature binding account.
This PDF is available to Subscribers Only