Abstract
Reading in peripheral vision is slow and inefficient. One plausible explanation for slow peripheral reading is that the peripheral retina lacks the chance to practice, unlike the fovea that is used constantly to read and thus is highly practiced. If so, then adequate training on a task that is novel to both the fovea and the periphery should yield the same magnitude of improvement at both locations. To test this “lack of practice” hypothesis, I tracked the improvements in contrast thresholds for identifying letters of a new alphabet-set, created by rotating each individual Greek letters by 180°, in the fovea and at 10° in the lower visual field. Letters were embedded within static two-dimensional Gaussian luminance noise. Six levels of Gaussian noise contrast were used (rms contrast: 0 to 0.2). Eighteen normally sighted observers were randomly assigned to one of 3 groups: training at the fovea, training at 10° lower field and a no-training control group. A pre-test consisted of measuring contrast thresholds for identifying the rotated-Greek letters at the fovea and 10° lower field, for all noise levels. Letter size used was twice the acuity at each eccentricity. During training (over 5 consecutive days), contrast thresholds were measured at the trained eccentricity only. The post-test, identical to the pretest, was conducted on the day after training concluded. The control group received only the pre- and post-tests. After 5 days of training, observers of both training groups showed an average improvement of 25% in the optimal contrast thresholds for identifying rotated-Greek letters at the trained eccentricity and about 10% at the untrained eccentricity. The average improvement for the control group between pre- and post-tests was also about 10%, at both eccentricities. The similar magnitude of improvement at the fovea and 10° lower field following training suggests that the periphery is capable of learning as well as the fovea.
Supported by NIH grant EY12810.