アイテム詳細

要旨

Fusion of dierent cues can improve the reliability of perceptual estimates. E.g., a more accurate size estimate can be achieved when visual and haptic size information is used in combination (Ernst Banks, 2002, Nature, 429-433). To combine dierent cues the observer has to know which cues belong together (i.e., the correspondence problem has to be solved). Here we examine whether subjects can learn to fuse two arbitrary cues based on their statistical co-occurence. To this end, we trained 6 subjects for about an hour with highly correlated stimuli that are usually uncorrelated, the luminance of an object (visual cue) and its stiness (haptic cue). To test for learning, we measured subjects' discrimination performances before (pre-test) and after (post-test) training. The discrimination task had two intermixed conditions: One condition in which the cues were consistent with the correlation during training (congruent) and another condition in which they were anti-correlated (incongruent). If training had an eect we would predict that the stimuli with congruent cues elicit an improvement in discrimination performance relative to the incongruent condition, because if the cues were truly fused after training the variance of the combined estimate should be lower. In agreement with our prediction we found a signicant interaction between pre- and post-test for the two congruent and incongruent conditions (F[1,5]=20,3; p<0.01). This indicates that subjects indeed picked up the correlation in the training phase and fused the two cues. We conclude that fusion of cues can be learned on a relatively short time-frame based on the statistics of their co-occurrence.