Extinction of Conditioned Flavor Preferences

Although noted as a proponent of associative learning theory, Bob Rescorla acknowledged that other mechanisms might be responsible for the within-event learning produced when two stimuli co-occur. To investigate this possibility, he conducted experiments in which rats experienced a compound of a novel flavor and a palatable nutrient, and demonstrated that a preference for the flavor established by this training did not show the pattern of extinction that might be expected of a preference based on a flavor-nutrient association. A review is presented of subsequent work on the extinction of such conditioned flavor preferences in rats. The results are found to depend on the motivational state of the rat in training and on test, on the match between the procedures on training and test, and on the details of the test procedure (the nature of the choice offered to the rat). When conditions are arranged appropriately, the extinction effect (a loss of the conditioned response) expected by standard associative theory can be obtained. What remains a problem for this theory is the observation (made originally by Rescorla himself) is that the effects of extinguishing a conditioned flavor preference are remarkably persistent. The failure to obtain recovery from the effects of the extinction procedure remains as a signal that this form of learning may involve processes other than the association formation used to explain many other forms of learning.