Abstract
The ability to mentally rotate complex images such as models of mechanical, chemical or biological structures is crucial for success in science and engineering. Our recent work suggests that the capacity of mental rotation is severely limited for updating local features such as colors and shapes of parts (Xu & Franconeri, 2015). We hypothesized that local-feature updating may be facilitated by providing appropriate motion cues during mental rotation. Observers were shown a plus-shape (400ms) with four distinctly colored arms, and asked to imagine the shape rotate clockwise by 90° during a short interval (120ms). During this interval, the screen was either blank, or it unpredictably depicted various rotation animations of the black outline of the plus-shape. Observers learned about the animations before the experiment and were instructed that the animations would be irrelevant. The plus-shape reappeared following the mental-rotation interval and observers determined whether it reflected their rotated mental image or two colors had been swapped (50% of trials). Trials where no color-swap occurred allowed the most straightforward test of whether the depicted image matched the observers' mental image of a 90° clockwise rotation, while the color-swapped trials provided a check for response bias. Relative to the blank-interval condition, the identification of the correctly rotated images was improved by a clockwise-rotating animation, hurt by a counterclockwise-rotating animation, and unaffected by other control animations. The rejection of the color-swapped images was unaffected by any of the animations, ruling out response bias as an explanation. Because observers were clearly instructed that animations may be misleading (and they reported trying to ignore the animations), the motion of the irrelevant animation appears to influence feature location updating in an automatic fashion. The presence of motion information may spontaneously 'drag' features to their new locations during mental rotation.
Meeting abstract presented at VSS 2017