Academic journal article Cognitive, Affective and Behavioral Neuroscience

Two Types of Image Generation: Evidence from PET

Academic journal article Cognitive, Affective and Behavioral Neuroscience

Two Types of Image Generation: Evidence from PET

Article excerpt

Is there more than one method whereby people can generate visual mental images? Participants generated images after learning patterns in two ways. In one condition, they memorized descriptions of how segments are arranged to form patterns; in another, they memorized segments and mentally amalgamated them into patterns. In both conditions, identical stimuli cued them to form images while brain activation was monitored using PET. Comparison of the two imagery conditions revealed different activation between hemispheres when images were formed after patterns were learned by mentally combining segments versus when images were formed from stored verbal descriptions. Thus, images can be generated in at least two ways. However, this laterality difference was subtle; the majority of brain areas were activated in common across conditions. This overall similarity is counter to what would be expected if image generation is simply perceptual exploration in the absence of appropriate stimuli, as is posited by perceptual activity theory.

One of the major advances of cognitive neuroscience is that mental functions that previously were regarded as unitary and undifferentiated have been shown to have a complex underlying structure. Visual mental imagery is one such function. Previous research has indicated that it is possible to decompose imagery into various processes, such as image generation (the process of creating an image on the basis of stored information), image inspection (interpreting patterns in images), and image transformation (changing the appearance of visualized patterns; see, e.g., Farah, 1984; Kosslyn, 1994; Kosslyn, Brunn, Cave, & Wallach, 1984; Kosslyn et al., 2004). In this article, we push this line of research one step further, and show that even a single process-image generation-can be further decomposed. We use positron emission tomography (PET) to provide evidence that visual mental images can be generated using two distinct methods.

The process of generating images is constrained in part by the nature of the information that is used to form the images. People typically look at several portions of an object when examining it, which implies that separate representations of parts and characteristics of objects are encoded. If a visual mental image is later to be formed, the stored parts and characteristics must be properly amalgamated. Kosslyn, Maljkovic, Hamilton, Horwitz, and Thompson (1995) have argued that there are two ways in which parts can be arranged to form images. On the one hand, they can be arranged by using categorical spatial relations, which specify equivalence classes (e.g., above); on the other hand, they can be arranged by using metric information that specifies (with precise distances) exactly how parts are juxtaposed (see also Kosslyn, Thompson, Gitelman, & Alpert, 1998). Kosslyn et al. (1989; Kosslyn, Maljkovic, et al., 1995; Kosslyn et al., 1998) showed that categorical spatial relations are encoded more effectively in the left cerebral hemisphere, whereas coordinate spatial relations (which specify metric information relative to an origin) are encoded more effectively in the right cerebral hemisphere; this dissociation provides evidence that the two types of spatial relations are encoded by distinct mechanisms (see also Chabris & Kosslyn, 1998; Hellige & Michimata, 1989; Jager & Postma, 2003; Kosslyn, Chabris, Marsolek, & Koenig, 1992; Laeng, 1994; Laeng, Chabris, & Kosslyn, 2002; Slotnick, Moo, Tesoro, & Hart, 2001). If two types of spatial information are encoded and stored, in principle, each type could be used subsequently to reconstruct the spatial arrangement of the parts in a mental image; and if different processes are required to use the two types of spatial representations, we should find some brain areas more active during one type of processing and others more active during the other type of processing (see Kosslyn, 1987, 1994).

Kosslyn, Maljkovic, et al. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.