Mental imagery interferes with perception. This, an example of the 'Perky effect', was studied for vernier acuity. Mean accuracy for reporting the offset of vertical line targets declined from 80% to 65% when subjects were requested to imagine vertical lines near fixation. Images of horizontal lines or of a grey mist in the fixation region lowered accuracy to a similar extent. However, accuracy was barely affected when the image was requested 1.5 deg or more from the target. The Perky effect remained strong for at least 4 s after an instruction to 'clear' the image away. The results were not due to imagery-induced changes in fixation, pupil diameter, or accommodation, or (at least primarily) to central attentional or decisional factors. Rather, imagery produces a local, pattern-insensitive, and relatively long-lasting reduction in visual sensitivity. The sensitivity loss may be mimicked by a 0.24 log unit reduction in target energy.

Mental visual imagery interferes with vision: the Perky (1910) effect. Is the effect optical, sensory, perceptual, attentional, or just a response bias? Acuity was measured (in undergraduates and graduates) using target lines, with and without images (of lines). Optics (fixation, pupil size, accommodation), response bias, global attention (effort; diversion of attention to imagery), perceptual assimilation (target incorporation by imagery) and perceptual masking (of target by imagery) all fail to explain the effect. Foveally, local attention plays a limited role, as the Perky effect in divided attention is half that in focused attention, but this interaction vanishes with extrafoveal targets. Images produce primarily sensory interference, mimicking a reduction in target energy.

Healthy subjects demonstrate leftward bias on visual-spatial tasks. However, young controls may also be left-biased when drawing communicatively, depicting the subject of a sentence leftward on a page relative to the sentence object, that is, a spatial-syntactic, implicit task. A leftward visual-spatial bias may decrease with aging, as right-hemisphere, dorsal, visual-spatial activation may be reduced in elderly subjects performing these tasks. We compared horizontal and radial (near-far) visual spatial bias, and spatial-syntactic bias, in healthy young and aged participants. Both horizontal and radial visual-spatial bias were smaller in aged participants when explicitly, but not implicitly assessed. Mean implicit far bias was greater in aged subjects, although this varied by task. We observed less implicit, spatial-syntactic left bias in aged than young participants. These results may be consistent with relatively less dominance of right hemisphere, dorsal spatial systems with aging. (JINS, 2008, 14, 562-570.)

Imagery interferes with visual acuity (the "Perky effect") when an image is close to a visual target and both the image and the acuity target are located in the same depth plane. Whether imagery-induced interference occurs when a mental image and a target are separated by induced depth was investigated. Participants projects an image in front of or behind a vernier acuity target on a frontal or back plan suggested by the panels of an outline cube. A drop in accuracy for the target was found when an image was projected in front of, but not behind, the target. Thus, induced depth can influence the Perky effect. By contrast, real lines interfered with the target regardless of perceived depth plane, which is inconsistent with the hypothesis that imagery and perception are equivalent. Results support the hypothesis that images interfere with perception only when the participant must see through an image to obtain information specifying the visual target.

Todetermine whether self-generated visualimagery alters likingratings of merelyexposed stimuli, 79 college students were repeatedly exposed to the ambiguous duck-rabbit figure. Half the participants were told to picturethe imageas a duck and halfto picture it as a rabbit.Whenparticipants made liking ratings of both disambiguated versions of the figure, they rated the version consistent with earlier encodingmore positively than the alternate version. Implicationsof these findings for theoretical modelsof the exposure effect are discussed.