Abstract
Previously, we reported that following five days of intensive practice at the fovea or 10° eccentricity, observers improved in their contrast thresholds for identifying unfamiliar letters at their trained retinal location (Chung, 2002 VSS). However, these improvements did not transfer from the fovea to 10° eccentricity, or vice versa. Here, we investigated whether similar learning effects would transfer across untrained peripheral retinal locations. We measured contrast thresholds for identifying rotated-Greek letters, as in Chung (2002), at 10°, 15° and 20° eccentricity in the lower visual field before and after five consecutive days of training. Letters were twice the acuity at each eccentricity, and were embedded in the presence of five levels of external static white noise (rms contrast: 0.03 to 0.2). Twenty-four observers with normal vision were randomly assigned to one of four groups: training at 10°, 15° or 20° in the lower field, and a control group. A pre-test consisted of identifying the rotated-Greek letters at each of the three eccentricities, for all noise levels. During training, observers in the training groups performed the same letter identification task at their assigned trained retinal location; whereas observers in the control group were trained on an unrelated Vernier judgment task. A post-test, identical to the pre-test, was conducted after training concluded. Following training, observers in the three training groups showed very similar magnitudes of improvement in performance at their trained location (e.g. the sampling efficiency increased by 2.3 to 2.5x). These improvements transferred almost completely to the other two untrained retinal locations (the sampling efficiency increased by 2.1 to 2.4x). In comparison, the sampling efficiency of the control group observers increased by only 1.4x. Unlike our previous results between the fovea and the periphery, the current finding indicates that improvements that occur at a peripheral retinal location can be transferred to other untrained peripheral retinal locations. This finding has significant implication for the rehabilitation of patients with central vision loss.
Supported by NIH grants T35-EY07088 and R01-EY12810.