Scaling of Size and Contrast Equalises Extrafoveal Word Recognition at All Word Lengths

Abstract

Abstract: : Purpose:Contrast sensitivity for letter recognition can be equalised across eccentricities by scaling letter size and contrast according to local spatial and contrast scales of the underlying visual mechanisms. However, an incommensurable deficit has been reported for extrafoveal reading. It has been attributed to reduced visual span. In these studies contrast was not scaled. Hence, extrafoveal presentations had lower effective contrast, which might explain poorer performance/shorter span. Our «double scaling’ hypothesis predicts that plots of log contrast sensitivity versus log stimulus size should differ only quantitatively across retinal eccentricities in word recognition, too. Methods:We measured contrast sensitivities corresponding to 59% correct word recognition from a set of 8 known words as a function of size (0.28-9.0 deg) across eccentricities (E=0-10 deg) for four different word lengths (1, 3, 5 and 7 letters plus a flanking letter «H’ at each word end). Results:Irespective of word length extrafoveal recognition was the same as that of the fovea when extrafoveal letters were bigger and had higher contrast. Sensitivity curves had the same shape at all eccentricities but the curves were shifted along logarithmic size and contrast axes, in agreement with the double scaling hypothesis. Moreover, the scaling of size and contrast required with increasing eccentricity was the same at all word lengths. Conclusion:The finding that scaling did not depend on word length means that visual span remains constant across eccentricities when extrafoveal stimuli have sufficient size and contrast. Thus, extrafoveal reading may be equalised with that of the fovea by band-pass filtering the letters/words to create room for the contrast increase required in the periphery.

Keywords: 539 reading • 586 spatial vision • 509 pattern vision