Abstract
Background. Visual perceptual learning (VPL) refers to improved performance following visual training, and is usually specific to trained retinal locations. However, covert spatial attention has been shown to generalize VPL to distant untrained locations in contrast sensitivity tasks. We investigated whether this finding generalizes to another basic dimension: visual acuity. We tested the effects of exogenous (involuntary) attention on VPL at both trained and untrained locations using an acuity task. Procedure. Two Landolt-squares were presented diagonally at isoeccentric locations. Each Landolt-square had a gap of varying size on the left or right side. A response cue indicated which stimulus was the target. Participants reported whether the target square had a gap on the left or right side. On the Pre- and Post-tests (1st and 5th sessions, respectively), performance was measured for the two diagonal pairs. During the intermediate Training sessions (2nd, 3rd and 4th sessions), only locations at one diagonal were tested (trained locations). Observers were divided into two training groups: a Neutral group, for which a central pre-cue was presented at fixation, and an Attention group, for which a peripheral pre-cue appeared adjacent to the upcoming target location. Both testing sessions employed neutral cues and were identical for the two groups. We derived psychometric functions and calculated 75% contrast thresholds. VPL at trained and untrained locations was assessed by changes between Pre- and Post-tests in overall performance and gap-size 75%-threshold. Results. We found a standard VPL pattern for the Neutral group: performance improved and threshold decreased more at trained than untrained locations-location specificity. However, for the Attention group, performance improved and threshold decreased at both trained and untrained location–no location specificity. Conclusion. Exogenous attention generalizes VPL to untrained locations in an acuity task, providing converging evidence that attention helps generalize learning.
Meeting abstract presented at VSS 2014