Odor qualities influence odor identification and naming performance in older adults
Poster presentation
Odor identification (with word cues) and naming (without cues) abilities are often impaired in aging adults. Severe impairment predicts future cognitive impairment, and clinical onset of dementia. Understanding the unique processing demands of the odor identification task is thus of high priority - however, investigations of this kind are sparse. We examined how olfactory-perceptual features that vary among the individual odors in the set, can be used to predict identification and naming performance in older adults. We used data from the Swedish National Study on Aging and Care in Kungsholmen (SNAC-K), where 2479 individuals (age 60-100 years) were assessed for odor identification and naming abilities using the Sniffin’ TOM test (a validated, slightly modified version of the original Sniffin’ Sticks identification test). In order to derive information about the perceptual differences between the 16 odors, we conducted a psychophysical rating experiment where 37 adult participants rated the pleasantness, intensity, familiarity and edibility of the odors. We also collected pairwise similarity ratings to establish the relative distinctiveness of the 16 odor qualities. Random effect logistic regression modelling was then used to predict the influence of the perceptual features on odor identification. Results show that the perceived odor intensity was the strongest predictor of identification success in the aging sample. Intensity also correlated significantly with naming and identification ability. Further, in agreement with previous research, unpleasant odors were more easy to identify than pleasant odors - pleasantness was the strongest predictor for naming ability. We conclude that for aging persons, odor identification and naming ability can in part be explained by the perceptual features of the odor. Our research approach can be used to optimize olfactory tests for different clinical conditions. This research was funded by the Swedish e-Science Research Center.