Abstract
Object recognition is facilitated after manual exploration relative to purely visual processing. Children and adults have been shown to focus on specific viewpoints of object structure during manual exploration. Research has stressed that manual exploration allows appropriate guidance of perceived visual information rather than a facilitation based on haptic information. The current experiment further assessed how visual and haptic object information influences performance on an old-new object recognition test. We compared manual exploration of 3D object with: passive viewing of a yoked video from manual exploration (visual information only), watching 2D information of their own manual actions on the objects (via computer screen), viewing 2D information from another participant (computer screen) while manually exploring objects or manual exploration with no visual information. After interacting with objects in one of the exploration study conditions, participants completed an object recognition task that incorporated several views of the studied objects, and unstudied objects. We predicted that performance would differ not only between groups due to different access of haptic and visual information, but also between planar and ¾ angle images of objects on the test. Results demonstrated that the combination of haptic and virtual visual information did not facilitate better recognition than either the passive or active conditions. Across measures of accuracy, response time, and sensitivity, the active condition group (with direct visual and haptic information) had higher performance than all other groups, while the haptic only group had the poorest performance. The three conditions with haptic exploration and 2D visual information did not differ from one another. The ¾ views and side view angles were recognized with higher performance for all study groups, compared to the front and back views. Our findings suggest that haptic interaction with objects does not facilitate performance unless coupled with visual information from actual 3D objects.