Learning to Smell: Olfactory Perception from Neurobiology to Behavior

7. Odor Memory

7 Odor Memory To this point, we have been concerned with the role of memory in odor object recognition and hedonics. A whole range of other information-processing tasks exists that rely on memory, and the purpose of this chapter is to examine and interconnect them with the object recognition process that we have alluded to in the preceding chapters. We begin with an exploration of human olfactory cognition, including short- and long-term memory, paired associate learning, odor priming, and imagery. We then explore many of these same issues in the animal literature and explore underlying neural correlates. Olfactory Cognition in Humans An important question one might ask is whether there is any need to attempt an integration of olfactory cognition, odor memory, and object recognition, and thus we start this section by explaining why it is necessary. Higher cognitive processes clearly influence perception—top-down processing—as we briefly described in the preceding chapter. For example, being able to name an odor appears to affect all three aspects of olfactory perception; a nameable odor appears more intense (Distel and Hudson 2001), it evokes a more potent emotional response (Herz and Von Clef 2001), and it is better discriminated too (Rabin 1988). Such findings can be conceived in two basic ways, either as perceptual or nonperceptual phenomena. Both alternatives might operate together or alone, and each could be envisaged in several different forms. Perceptual accounts might involve (1) the direct activation of an odor object encoding ; (2) the refining of a representation, where the generation of a name for example, acts reciprocally to repress odor objects that are similar to the target’s encoding; and (3) the enhancement of the target’s level of activation. Nonperceptual accounts might involve (1) the recruitment of a second mode of representation (verbal), thus improving performance on memory and discrimination tasks (c.f., dual-coding; Paivio 1991); and (2) the alteration of the verbal description of a perceptual experience to bring it into line with information from the other senses. Whichever may be correct—and they are not mutually exclusive—top-down processing undoubtedly occurs (see earlier chapters) and we therefore need to know how the object recognition process that we have described is affected by it. A second reason for integrating the process of object recognition with findings from olfactory cognition is that many cognitive processes in the visual system, for example, directly utilize perceptual pathways. The most notable of these is the capacity for imagery (Kosslyn and Thompson 2003). In the visual system, imagery calls upon the integration of a range of processes, most notably retrieval of perceptual encodings from long-term visual memory and their instantiation in short-term memory (Baddeley and Andrade 2000; Kosslyn , Ganis, and Thompson 2003). Important conceptual similarities exist between imagery and top-down processing, even though these are often treated as separate topics in the literature (Cain and Algom 1997). If imagery is conceived as a perceptual process, which many findings from visual neuroscience suggest (Farah 2000), then the generation of an image when an odor’s name is provided, for example (imagine lemon), may involve virtually identical processes to those involved in top-down processing, where a name may act to refine or alter an extant perceptual experience. In other words the ultimate result of top-down processing may in certain cases be a perceptlike image . Two aspects of this are particularly important. First, if imagery is not a perceptual process as some have argued (Pylyshn 2003), this could imply that top-down effects are nonperceptual too (note the direct parallel with the discussion in the preceding paragraph). Second, several prominent olfactory theorists, especially Engen (1987), have argued that the olfactory system is incapable of imagery, whilst others (e.g., Cain and Algom 1997) suggest that imagery plays a key role in many olfactory tasks. Not only then do we need to understand how imagery (and top-down processing) may relate to object recognition, we also need to determine whether it constrains the type of exOdor Memory 189 planation that we can offer for imagery and top-down processing. In essence does an object recognition account of olfactory perception suggest that images are perceptual or nonperceptual phenomena? A third, and perhaps the most compelling, reason is that to date there is no account in the literature that presents a modal model of olfactory information processing. Rather, thinking in...