Higher-order processes in auditory-change detection.

seen i n pol ar projection (the configuration that normally elicits the most rigid percept) would be seen as deforming. Hence Si nha and Poggio's results clearly demonstrate that future psycho-physical investigation on the perception of 3D shapes will have to take into account learning processes that can take pl ace on relatively short time scales More generally they also l ead to reconsider classical schemes of 3D shape perception i n terms of: (I) the type of object representation involved i n visual processes, and (2) the existence of top-down control of these processes. The interpretation of a moving retinal image (1996) The dominance of static depth cues over motion parallax in the perception of surface orientation Perception 25 (suppl.), 40 Response from Sinha and Poggio L amouret, Cornilleau-Peres and Droul ez raise a number of very interesting points i n their Comment article. We shoul d like to take this opportunity to emphasize one of these issues that we find particularly important and intriguing but coul d not dwell upon adequately i n the original paper, for lack of space. Lamouret et al. remark on how learning initially depends on bottom-up sensory-information processing that uses generic biases such as those favoring object rigidity. However, the learning subsequently can overwhelm the results of such bottom-up processing. The percept, apparently, is controlled to different extents at different times by the generically processed sensory information on the one hand and object-specific l earned expectations on the other. The big question is: How does the brai n strike a compromise between sensation and, for want of a better term, hallucination? The parameters determining the relative contributions of the two quantities to the overall percept are likely to be a function of time i n two ways. (1) Expectations will exercise greater control i n determining percepts the longer the training time. (2) The bottom-up sensory information will become increasingly evident the greater the stimulus inspection time. The well-known hollow-mask illusion serves as a nice illustration of this point. The illusion often persists even under binocular viewing. If one subscribes to the accounts of the illusion that are based on familiarity, then it is reasonabl e to suggest that the greater the familiarity of an observer with faces, the more susceptible the observer will be to perceiving the illusion. On the other hand, the longer one binocularly inspects the hollow mask, the more …