Using a temporal integration task, subjects in 5 experiments were expected to combine information from temporally separated visual presentations. Evidence from these experiments indicated that perceptual information can be integrated with previously generated and currently maintained visual images to form a representation that contains information from each source. Properties and limitations of this integration process were also explored, including the time required to generated the image, the speed at which percepts were integrated with images, and the capacity of the representation. Implications for theories of visual processing and memory are discussed.Phenomenologically, visual images and visual percepts are alike in many respects.1 This apparent similarity has led to a great deal of research assessing the parallels between images and percepts by searching for common structural properties of percept-based and image-based representations as well as shared neurological substrates between the visual perception and imagery systems. For example, considerable research has demonstrated that both images and percepts encode spatial dimensions and spatial relations among objects (e.g., Farah, 1985;Finke & Pinker, 1982;Kosslyn, 1973Kosslyn, , 1978Kosslyn, Ball, & Reiser, 1978;Pinker, 1980;Pinker & Kosslyn, 1978;Weber & Harnish, 1974). In addition, decision times for comparison of two imagined objects correspond to those for two perceived objects (e.g., Moyer & Bayer, 1976;Paivio, 1975). Like visual perception, visual images are limited in spatial resolution in that the amount of information contained in an image is constrained by the imagined object's size and distance (e.g., Kosslyn, 1975Kosslyn, , 1976Kosslyn, , 1978Weber & Malmstrom, 1979) as well as eccentricity in the imagined visual field (e.g., Finke & Kosslyn, 1980;Finke & Kurtzman, 1981b). Furthermore, just as closer visual inspection of an object from different points of view reveals previously unrecognized properties of the object, there is some evidence that manipulation of mental images can also reveal patterns in the image not apparent when the image was originally formed (e.g., Brandimonte, Hitch, & Bishop, 1992;Pinker & Finke, 1980), although it appears that such insight is more difficult with images than with percepts (e.g., Chambers & Reisberg, 1985).The behavioral findings that suggest that image-based and percept-based representations share a common structure are complemented by research that has shown that visual imagery relies upon some of the same neural machinery as visual perception and generates similar patterns of neural activation.2 For example, using a radioactive marker to observe cerebral blood flow, Roland and Friberg (1985) determined that imagining an object is associated with activation in the occipital, posterior superior parietal, and posterior inferior temporal lobes, all of which are involved in perceptual processing (see also Goldenberg et al., 1989). In addition, using positron emission tomography, Kosslyn and his colleagues (Ko...