Generalization and discrimination tasks yield concordant measures of perceived distance between odours and their binary mixtures in larval Drosophila

Similarity between odours is notoriously difficult to measure. Widely used behavioural approaches in insect olfaction research are cross-adaptation, masking, as well as associative tasks based on olfactory learning and the subsequent testing for how specific the established memory is. A concern with such memory-based approaches is that the learning process required to establish an odour memory may alter the way the odour is processed, such that measures of perception taken at the test are distorted. The present study was therefore designed to see whether behavioural judgements of perceptual distance are different for two different memory-based tasks, namely generalization and discrimination. We used odour-reward learning in larval Drosophila as a study case. In order to challenge the larvae's olfactory system, we chose to work with binary mixtures and their elements (1-octanol, n-amyl acetate, 3-octanol, benzaldehyde and hexyl acetate). We determined the perceptual distance between each mixture and its elements, first in a generalization task, and then in a discrimination task. It turns out that scores of perceptual distance are correlated between both tasks. A re-analysis of published studies looking at element-to-element perceptual distances in larval reward learning and in adult punishment learning confirms this result. We therefore suggest that across a given set of olfactory stimuli, associative training does not grossly alter the pattern of perceptual distances.