In two experiments, we used a temporal integration task to investigate visual mental images based on information in short-term memory or generated from information stored in long-term memory (LTM). We specifically asked whether the two sorts of images rely on depictive representations. If mental images rely on depictive representations, then it should be possible to combine mental images and visual percepts into a single representation that preserves the spatial layout of the display. To demonstrate this, participants were asked to generate mental images and then combine them with visual percepts of grids that were partially filled with different numbers of dots. Participants were asked to determine which cell remained empty when the two grids were combined. We contrasted predictions of propositional or verbal description theories with those of depictive theories, and report findings that support the claim that mental images—based on either short-term or LTM—depict information.
