Reisberg D
J Exp Psychol Hum Percept Perform. 1983 Dec;9(6):966-79. doi: 10.1037//0096-1523.9.6.966.
We argue that the requirement of task-general mental resources marks more "cognitive" kinds of perception. In Experiment 1, subjects are slower to discover the alternative organization of ambiguous stimuli if they are simultaneously doing mental arithmetic; in Experiment 2, latency to first reversal is increased by a concurrent memory load of seven digits. In Experiment 3, counting tasks also slow subjects solving the Street figures and judging whether drawings depict possible three-dimensional objects. There is no effect of distractor tasks on the perception of random-dot stereograms. Further, the data argue that the pattern of disruption cannot be attributed either to the unfamiliarity of the judgments or to their latency. We suggest that the distractor tasks are competing with the perceptual judgments for working memory and that perceptual judgments demand memory whenever a perceptual "solution" cannot be locally confirmed.