Abstract
Crowding can increase or decrease with additional flankers, depending on the grouping among stimulus components. Here, we examined how letter crowding changes with increasing number of flanking letters, and whether the change is modulated by letter complexity. In the experiment, stimuli were letter strings constructed using lowercase Courier font with a print size of 1° and a spacing of 1.16 × x-width. Identification accuracy was measured for the middle target letter which was flanked by either two or four flankers and presented at 10° to the left or right of fixation. Flankers were always duplicate letters, and either identical to or different from the target letter. There were a total of 676 target-flanker combinations plus 26 isolated letters. Identification accuracy was nearly perfect (0.98) for isolated letters. When increasing the number of flankers from two to four, no change in performance (0.68 vs. 0.75) was observed for identical target-flanker condition. For the non-identical condition, identification accuracy was lower for pentagrams (0.44) than for trigrams (0.55). Analysis of the error trials revealed that the deterioration resides in mislocation rather than crowding. There were more mislocation errors in pentagrams (0.24) than in trigrams (0.16). Linear regressions were performed to assess the effect of letter complexity. We found that higher flanker complexity led to worse performance while higher-complexity targets made identification easier. Specifically, more complex flankers induced greater crowding and more mislocation errors. Higher target complexity was associated with lower mislocation rates, and corresponded to less increase or even reduction in mislocation errors when doubling flankers. Our findings demonstrate that the letter crowding or the effect of grouping among stimulus components does not change with number of flankers. However, additional flankers do lead to more letter substitution, which is modulated by target letter complexity.
Meeting abstract presented at VSS 2016