Abstract
Purpose:
Reading speed in peripheral vision can be improved through perceptual learning. Perceptual learning paradigms typically require practicing a demanding task hundreds of trials per day, which can be an impediment to compliance from patients. To minimize training effort, Yu (ARVO 2013) developed a non-task-based training procedure utilizing stimulus exposure and identity priming. The goal of this study is to identify the factor(s) that determine the effectiveness of non-task-based learning in improving peripheral reading speed. The findings will be of great importance for developing a useful form of low-vision reading rehabilitation.
Methods:
Twenty-eight normally-sighted young adults were randomly assigned to a no-training control group or one of four training groups that differed with respect to the inclusion of different design components. Pre- and post-tests consisted of measurements of visual-span sizes (number of letters recognized without moving the eyes) and reading speeds using the rapid serial visual presentation method at 10° above and below fixation. Training occurred at 10° below fixation. Three design components were visual exposure to crowded letter (target letter flanked by two other letters on the left and right), visual exposure to uncrowded letter (target letter alone), and identity priming (pre-knowing target letter identity). In five training sessions (1430 trials/session), group A viewed crowded letters, group B viewed crowded letters with identity priming, group C viewed uncrowded letters with identity priming, and group D viewed letters with identity priming under both crowded and uncrowded conditions.
Results:
Comparisons between different groups revealed an essential role of crowding in ensuring reading-related learning. Groups A, B and D showed significant improvement in visual-span size (gaining 1.2 to 1.4 letters; all p<0.0005). Only stimulus exposures that occurred simultaneously with identity priming enabled learning transfer to the untrained reading task (62% increase for group B and 78% for group D; all p<0.0005). Learning also transferred from the trained (lower) to the untrained (upper) visual field (all p<0.03).
Conclusions:
Non-task-based training can greatly reduce learning effort. Furthermore, including crowding and identity priming in the training paradigm is the prerequisite for effective learning in improving peripheral reading speed.