Decomposing the spatiotemporal signature in dynamic 3D object recognition.

The current study investigated the long-term representation of spatiotemporal signature (J. V. Stone, 1998) and its coding nature in a dynamic object recognition task. In Experiment 1, the observers' recognition performance was impaired by an overall reversal of the studied objects' learning view sequences even when they were unsmooth, suggesting that the spatiotemporal appearance of the objects was used for recognition, and this effect was not restricted to smooth motion condition. In another four experiments, a feature reversal paradigm was applied that only the global-scale or local-scale dynamic feature of the view sequences was reversed at a time. The reversal effect still held, but it was selective to the sequence's feature saliency, suggesting that statistical representation based on specific features instead of the whole view sequence was used for recognition. Furthermore, top-down regulation on sequence smoothness was observed that the observers perceived the objects as moving in a smoother manner than they actually were. These results extend an emerging framework that argues the spatiotemporal appearance of a dynamic object contributes to its recognition. The spatiotemporal signature might be coded in a feature-based manner under the law of perceptual organization, and the coding process is adaptive to variation of the sequence's temporal order.