Abstract
Background: Previous studies have reported reduced crowding for low contrast stimuli (Kothe & Regan, 1990: Optom Vis Sci; Simmers et al., 1999: Ophthal Physiol Opt). Both defined crowding as a change in acuity for single vs. flanked letters, and used foveal viewing. This study sought to confirm these findings using a more traditional crowding task (i.e. comparison of single and flanked letter identification at different retinal eccentricities). Methods: 10 observers (18–34 yo; VA 20/14) identified 0.34 deg letters and letters flanked by an ‘x’ on either side (xax) at fixation and ±1, 3, 5, and 7 letter spaces (±0.34, ±1.02, ±1.70, and ±2.38 deg) from fixation on the horizontal meridian. All 25 letters (other than x) were presented at each location; letter and position were randomly selected from trial-to-trial. In separate blocks, the stimuli were ∼100, 30, and 10% contrast. Crowding was defined as the difference between single letter and flanked letter performance. Results: There was little difference in crowding for the three contrasts. Averaged across position, there was 14±0.3, 15±0.1, and 19±0.2% crowding for the 100, 30, and 10% contrasts, respectively [F(2, 18) = 2.3, p = 0.127]. There was little change from 100% contrast for single or flanked letters presented with 30% contrast, but with 10% contrast, performance was significantly reduced for both stimulus types. Accuracy was similar across positions for single letters presented with 10% contrast and flanked letters presented with 100% contrast. Conclusions: Unlike previous reports, these data show no change in crowding with reduced contrast stimuli. It is possible that 10% is not sufficiently reduced for crowding to disappear, but Kothe and Regan showed reduced crowding for 11% contrast stimuli. The data suggest that the effect of flanking letters may be to reduce the effective contrast of the target letters.