I’ve often perceived two or more images in my visual plane get mixed up, even if they are pretty much apart. The same holds good for words: reading about energy issues and encountering the word ‘green’ a lot, and later any word that starts with ‘gr’ (say greek) often is read as green. This sort of false ‘binding’ of images/words is the main research of the Princeton psychologist, Anne Treisman. Her Princeton page reads:

“…when we are visually overloaded, the binding of features can fail. For example, when research participants are given less than a second to attend to an array of images that include the letter "S" as well as straight lines nearby it, they often report seeing dollar signs, where no dollar signs were present.

“To explain these observations, Treisman proposes that sensory stimuli are initially registered both in separate feature maps and in a location map. We recombine the features by focusing on one object at a time and form a temporary representation, or object token, of the features currently in the attended location. We can then compare the object token to a store of knowledge built up over time and identify the object as, for example, a cat, or a pencil, or a telephone. The object token tells us what it looks like at this moment of time and updates its position and appearance if it moves and changes. It also allows us to represent novel objects that we have never seen before and therefore cannot identify. We seem to be limited to holding three or four object tokens at a time in temporary memory, and we lose the bindings even for those if attention is attracted elsewhere.

"Treisman studied a patient with Balint’s syndrome, a disease characterized by lesions to the parietal cortex, which destroy the ability to localize objects in space. Treisman and her collegues predicted that this would also lead to failures of object binding and confirmed that the patient did indeed see illusory combinations of features, even with presentations of just two colored letters for up to ten seconds."