Flavour–nutrient learning in humans: An elusive phenomenon?

Abstract

One widely cited model of how humans acquire liking for different foods is flavour–nutrient learning, where associations between the orosensory properties of the ingested food or drink (the flavour CS) and positive consequences of nutrient ingestion (the UCS) lead to acquired liking for the flavour (flavour–nutrient hedonic learning: FNL-H). Likewise, an association between the CS and the post-ingestive effects of ingested nutrients has been suggested to lead to learning about how satiating a particular food is (flavour–nutrient satiety learning: FNL-S). However, whereas there is evidence for both FNL-H and FNL-S in experimental studies with non-human animals, evidence in humans is less convincing, with many failures to find the predicted changes in liking, preference or intake following repeated flavour–nutrient pairings. The present short review considers how subtle differences in experimental design might underlie this inconsistency, and identifies key design features which appear to increase the likelihood of success in human flavour–nutrient learning studies. Key factors include CS novelty, the level of nutrients ingested during training, the appetitive state of the consumer and individual consumer characteristics. A further complication is competition between FNL-H and FNL-S, and with other associations such as flavour–flavour learning. From this it is possible to make important inferences about the nature of human flavour–nutrient learning which firstly suggest that it has important similarities to that seen in other species, but secondly that the laboratory investigations of both FNL-H and FNL-S in humans can be compromised by subtle but important variations in experimental design.