Specificity of learning curvature, orientation, and vernier discriminations

Training significantly improves the performance of many perceptual tasks. Different visual tasks share some "front-end" neuronal mechanisms but rely (partly) on different neuronal mechanisms for further analysis. Perceptual learning might occur on the early common levels of visual information processing or else on the later, more specialized levels. Eighteen observers trained in three visual hyperacuity tasks, namely curvature, orientation, and vernier discrimination that probably share a common first stage of analysis based on detection of oriented line elements. Speed of improvement did not differ significantly between these tasks. There was no transfer of improvement from one task to another, indicating that the neuronal mechanisms underlying the three tasks are at least partly non-identical and that learning does not take place on the first common levels of analysis. This result constrains the possible localizations, in the human brain, of perceptual learning. The study also demonstrates that perceptual learning can be used as a tool to increase our knowledge on the sequence of operations during (visual) pattern recognition.