Location Specificity in Perceptual Learning: A Revisit

C. Yu,S. Klein,D. Levi
DOI: https://doi.org/10.1167/7.9.87
2010-01-01
Journal of Vision
Abstract:Perceptual learning (PL) is specific to the practiced retinal location, a signature property in many PL tasks. We investigated this signature with a contrast discrimination learning task. Contrast learning for a V-Gabor in the lower-right quadrant (Lower-Loc) of the visual field was not transferred to the upper-right quadrant (Upper-Loc) for the same V-Gabor, replicating location specificity. However, simultaneous training of a V-Gabor in Lower-Loc and a H-Gabor in Upper-Loc greatly improved contrast discrimination of a V-Gabor in Upper-Loc. To separate the contributions of Lower-Loc V-Gabor training and Upper-Loc H-Gabor training, we conducted (a) Lower-Loc V-Gabor training followed by Upper-Loc H-Gabor training and (b) same training in a reversed order. In (a) initial Lower-Loc V-Gabor training had little effect on Upper-Loc V-Gabor discrimination, but later Upper-Loc H-Gabor training greatly improved Upper-Loc V-Gabor discrimination. In (b) initial Upper-Loc H-Gabor training also greatly improved Upper-Loc V-Gabor discrimination, but later Lower-Loc V-Gabor training further improved Upper-Loc V-Gabor discrimination. These data indicated that improved Upper-Loc V-Gabor discrimination mainly depended on training in the same retinal location, even with an orthogonal stimulus. However, after a location had been trained or sensitized, learning of the same stimulus in a different location could then be transferred to the sensitized location. A control experiment demonstrated that even simultaneous training of V-Gabor contrast discrimination in Lower-Loc and 45°-Gabor orientation discrimination, which was a completely different task, in Upper-Loc also greatly improved Upper-Loc V-Gabor contrast discrimination, indicating that training-induced location sensitization might be task unspecific. We conclude that PL might include a location-specific but stimulus-or-task-unspecific process, and a stimulus-specific but location-unspecific process. Contrast learning in the periphery is mainly determined by location specific training of probably any task. In a less degree it is also affected by location-unspecific stimulus training, provided that the target location has been sensitized.
What problem does this paper attempt to address?