Abstract
Several studies have shown that humans encode dynamic information such as rigid (e.g., rotations in depth) and non-rigid motion (e.g., facial expressions) for recognition purposes. Our goals in the present study were to extend previous work by using familiar objects which had a wide range of shapes and to determine possible influences of shape transformations on the encoding of dynamic information. We therefore tested the role of non-rigid shape changes in recognizing common animals and objects. To create these changes, we morphed between members of the same category (e.g., German shepherd and dachshund from the category dog). In a same/different categorization task, participants were shown a continuous morph sequence, followed by a static test image, and had to decide whether the image was from the same category as the morph sequence. Importantly, this paradigm does not presuppose that participants perform shape transformations. We found that participants responded more quickly if test images were sampled towards the end of the morph sequences, which replicates previous results with rigid rotations of objects and non-rigid facial expressions. This dynamic benefit was not a recency effect because it was eliminated when we scrambled the frame order of the morph sequences while preserving the continuity of the first and last 150 ms of the sequences. Overall these findings indicate that dynamic information plays a role in the recognition of diverse familiar shapes, and suggest a general mechanism that encodes continuous shape transformations.