Purchase this article with an account.
O Ehrt, CB Williams, RF Hess; Detection- and discrimination thresholds for Landolt C in contour interaction . Invest. Ophthalmol. Vis. Sci. 2002;43(13):2837.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Purpose: Crowding and contour interaction are usually measured at the visual acuity limit as percent correct. We used a larger stimulus to see if these results generalize to stimuli closer to reading tasks and to be able to investigate the spatial frequency properties of contour interaction. Comparing contrast threshold for detecting and discriminating the stimulus could further help understand the mechanisms of contour interaction. Methods: The contrast thresholds for detecting the presence of a dark Landolt C (visual angle 0,5°) surrounded by high contrast bars at 0 - 2 gap width separation were determined on 2 normal subjects in a 2-interval AFC procedure with constant stimuli for foveal and 10° eccentric presentation (spatially scaled stimulus). Thresholds for discriminating the location of the gap were measured in the same conditions with a 4 - AFC procedure. We repeated these measurements with a filtered C (mean SF 1.9 cycles/letter) surrounded by filtered bars (mean SF 0.5 - 8 c/letter) at 0.5 bar width separation. Results: Mean contrast threshold increased from 0.07 (without bars) to 0.085 (bars abutting) for detection and from 0.11 to 0.17 for discrimination. The ratio c disc / c detect increased from 1.5 to 2.1 for dark bars and to 2.5 for white bars. Both contrast thresholds and their ratio did show a marked tuning function with regard to the spatial frequency of the bars. No qualitative difference was found between central and peripheral presentation. Conclusion: Discrimination and detection show different results. Discrimination is more affected by flanking bars than detection. Thus masking does only account for part of the - weak - contour interaction that is present in normal subjects.
This PDF is available to Subscribers Only