Immediate perception of a reward is distinct from the reward’s long-term salience

Abstract

Reward perception guides all aspects of animal behavior. However, the relationship between the perceived value of a reward, the latent value of a reward, and the behavioral response remains unclear. Here we report that, given a choice between two sweet and chemically similar sugars—L- and D-arabinose—Drosophila melanogaster prefers D- over L- arabinose, but forms long-term memories of L-arabinose more reliably. Behavioral assays indicate that L-arabinose-generated memories require sugar receptor Gr43a, and calcium imaging and electrophysiological recordings indicate that L- and D-arabinose differentially activate Gr43a-expressing neurons. We posit that the immediate valence of a reward is not always predictive of the long-term reinforcement value of that reward, and that a subset of sugar-sensing neurons may generate distinct representations of similar sugars, allowing for rapid assessment of the salient features of various sugar rewards and generation of reward-specific behaviors. However, how sensory neurons communicate information about L-arabinose quality and concentration—features relevant for long-term memory—remains unknown.DOI: http://dx.doi.org/10.7554/eLife.22283.001

Antifeedant activities of L-arabinose to caterpillars of the cotton bollwormHelicoverpa armigera(Hübner)

“…Our data show no specific L-arabinose-sensitive GRNs were found in the maxillary sensillum of H. armigera caterpillars, but the activities of GRNs to other substances, including GRNs sensitive sucrose, host leaf saps, and even the deterrent sinigrin, could be inhibited when combined with L-arabinose, suggesting the inhibition effect of L-arabinose is non-specific and not due to L-arabinose-sensitive GRNs in the maxillary sensilla. In contrast to findings in Drosophila where preference for L-arabinose is related to the presence of sweet gustatory receptor (McGinnis et al , 2016;Miyamoto et al , 2013), our electrophysiological and behavioral data suggest that L-arabinose is not a sweetener to H. armigera caterpillars. Activity of the deterrent sinigrin on sinigrin-sensitive GRNs was also inhibited when mixed with L-arabinose; therefore, we posit that L-arabinose could activate neither sweet GRNs nor bitter GRNs in the maxillary sensilla.…”

Antimicrobial peptides modulate long-term memory

“…Four hours after training mRNA was sequenced from 4 to 6 individual female fly heads and compared that to age matched untrained fed flies. To distinguish gene expression changes linked to CS-US association from gene expression changes due to starvation, or exposure to odor or sweet sugar, we also performed mRNA sequencing from three control groups; i) starved flies exposed to just octanol (CS only), ii) starved flies trained with octanol and L-sorbose (a sweet, non-nutritious sugar that produces robust short- but weak long-term memory [ 13 , 14 ] ( S1 Fig ) as US, and iii) an independent group of flies trained with sucrose as US after a month, to rule out the variation in gene expression in different population of flies ( S2 Table ). Expression of mRNA that changed only in both sucrose-trained groups compared to naive, CS alone or sorbose trained group, was deemed to be associated with long-term appetitive memory ( Fig 1B , S2 Fig and S2 Table ).…”

Regulated Intron Removal Integrates Motivational State and Experience

“…With the exception of L-arabinose, most sweet but non-nutritious sugars don’t produce robust long-term memory (Burke and Waddell, 2011; McGinnis et al, 2016). What is the implication of a sweet but non-nutritious sugar increasing spliced Orb2A transcript level?…”