Perceptual Learning Increases Sampling Efficiency
No Thumbnail Available
Although visual orientation discrimination has been shown to improve with training, the mechanism(s) underlying this task improvement remain poorly understood. One possibility that is investigated in this paper is that an increase in sub-sampling efficiency with practice causes a reduction in orientation discrimination thresholds. That is, a more reliable estimate of the stimulus orientation could be obtained by sampling a larger part of this stimulus. We used the classification image method (Ahumada, 1996), which relates the response of the observer to the orientation variability at different locations in the stimulus, to test whether a change in sub-sampling occurs with perceptual learning on an orientation discrimination task. The results showed that a decrease in orientation thresholds was accompanied by an increase in sub-sample size. An effect for eccentricity was observed; while the decision bias, the part of the stimulus that drives the decision, was restricted to the inner part of the stimulus before training, an outward spread of sampling with training was observed. Furthermore, a radial bias was observed after training, with the largest spread in sub-sampling observed for the top quadrant aligned with the stimulus orientation. These results showed an increase in sub-sampling efficiency with training, making it a likely mechanism to underlie perceptual learning.
Faculteit der Sociale Wetenschappen