Dissociation between visual attention and visual mental imagery

Visual mental imagery (which involves generating and transforming visual mental representations, i.e., seeing with the mind's eye) and visual attention appear to be distinct processes. However, some researchers have claimed that imagery effects can be explained by appeal to attention (and thus, that imagery is nothing more than a form of attention). In this study, we used a size manipulation to demonstrate that imagery and attention are distinct processes. We reasoned that if participants are asked to perform each function (imagery and attention) using stimuli of two different sizes (large and small), and that stimulus size affects the two functions differently, then we could conclude that imagery and attention are distinct cognitive processes. Our analyses showed that participants performed the imagery task with greater facility at a large size, whereas attention was performed more easily using smaller stimuli. This finding demonstrates that imagery and attention are distinct cognitive processes. Dissociation Between Visual Attention and Visual Mental Imagery Visual mental imagery involves creating, interpreting, and transforming visual internal representations (while "seeing with the mind's eye") whereas attention involves selecting some information for more detailed processing (while discarding other information). Thus, at first glance, the two functions appear distinct. Nevertheless, some researchers have claimed that mental imagery effects do not reflect characteristics of a distinct form of internal representation, but rather are best understood as products of attention. Pylyshyn (e.g., 1989, 2002, 2003), for example, has proposed that results from many "imagery" experiments actually reflect the allocation of attention to different portions of the space that would be occupied by an image. Pylyshyn (2002, p. 158) states that “...the use of visual indexes and focal attention provides a satisfactory explanation for how spatial properties are inherited from the observed scene, without any need to posit spatial properties of images.” For example, when visualizing a house on a blank wall, they may think “the front door would be where the speck on the wall is”, with attention's being allocated to different regions of space as the image is constructed descriptively. Similarly, Pani (2002) has echoed Pylyshyn's view that phenomena attributed to mental imagery (such as the visualization of letters within a 4x5-cell grid) are actually due to the allocation of attention to defined regions of space. This idea gains credence because Craver-Lemley and Reeves (1992) have shown that imagery and attention can interact in some circumstances: in particular, when attention is divided, the Perky effect (i.e., the interference with perceiving that is produced by a mental image occupying the same space as a foveal visual percept) is half of that observed when attention is focused. If the Perky effect is taken as evidence that imagery and perception rely partly on the same neural systems, then the fact that divided attention decreases the effect whereas focused attention increases it could suggest that imagery effects arise, at least in part, from focused visual attention. We should note, however, that this effect of attention on imagery/perception interference only occurs when the perceptual targets are in the periphery of the visual field. There is also increasing evidence that both visual imagery (e.g., Kosslyn & Thompson, 2003; Kosslyn, Thompson & Ganis, 2006) and visual attention can lead to increased activity in early visual cortex (e.g., Silver, Ress, & Heeger, 2005), and attention may enhance performance of tasks that rely on this neural structure (e.g., Hopfinger & West, 2006). These results make it difficult to disentangle effects due to imagery versus attention (but, see Offen, Schluppeck & Heeger [2009] for evidence that attention and visual short-term memory rely on different processes in early visual cortex). In addition, Grossberg (2000) has suggested that a combination of mismatched attentional and topdown expectancy effects can give rise to the experience of perceiving a stimulus in its absence in the case of schizophrenic hallucinations. However, Ishai, Haxby, and Ungerleider (2002) conducted a neuroimaging study in which they explicitly examined the effects of attention on imagery processing. They asked participants to visualize famous faces, and in one condition, asked them to focus their attention on a particular feature of the face (such as nose or lips). They found that requiring attention during imagery increased activation in some frontal and parietal regions (only a subset of areas activated by imagery), suggesting that the two processes are different. On the other hand, this result could be interpreted as showing that focusing attention to particular features represents higher resolution imagery to the target area, and so the study may not be considered a direct comparison of imagery and attention processes. If attention and imagery rely on distinct mental processes, then there should exist some variables that affect imagery and attention differently. The size at which people are asked to perform an attention or an imagery task may be one variable that affects each function differently. In particular, for attention, there is evidence that people have more difficulty detecting faint visual signals if their attention is distributed over a large area than a small one (Eriksen & St. James, 1986); conversely, for imagery, they have more difficulty detecting parts of a visualized object if the image is small than if the image is large (e.g., Kosslyn, 1975). Along the same lines, they also have poorer memory for objects visualized at small sizes versus large sizes (Kosslyn & Alper, 1977). In the present study, we compare directly the effects of size in imagery versus attention using matched paradigms with the same participants . If imagery and attention rely on the same processes, then manipulating the size of objects to be attended to or to be visualized should have the same effect. If, on the other hand, varying the size of the objects on which processing takes place produces different results for imagery and attention, then the two functions cannot rely on identical sorts of processing -and we are justified in concluding that imagery cannot be reduced to attention.