Blocking of acquisition of a taste aversion by a context experienced prior to the taste

Effects of extended context discrimination training and context extinction on transfer of context dependency of conditioned flavor aversion

We trained rats in a context discrimination paradigm by pairing a sucrose solution with lithium chloride in one context (conditioning context) and simple exposure to the same fluid in a second (neutral) context to establish a context-dependent aversion to the conditioned fluid. We then investigated whether transfer of the context dependency to a test fluid (a sodium chloride solution) was affected by two post-discrimination training treatments, an extended context discrimination training, and non-reinforced exposure to the conditioning context (context extinction). We found that the context-dependent flavor aversion that had been specific to sucrose transferred to the test fluid after the extensive training (Experiment 1). Context extinction eliminated the transfer effect that had been observed immediately after the context discrimination training (Experiment 2). In addition, an aversion acquired by sucrose through a simple conditioning of sucrose-LiCl pairings did not generalize to the test fluid (Experiment 3). These results emphasize the importance of a Pavlovian excitatory association between the conditioning context and nausea as a primary source of transfer of the context dependency, rather than a generalization of aversion acquired by the conditioned fluid to the test fluid.

Serial overshadowing of taste aversion learning by stimuli preceding the target taste

Three experiments tested whether events taking place before a rat has access to a target taste, sucrose, can proactively interfere with the acquisition of a sucrose aversion when sucrose is followed by a lithium chloride injection. Using a serial overshadowing procedure with various delays before lithium injection, proactive interference by a taste (Experiments 1 and 3) and by a novel context (Experiment 2) was found following two conditioning sessions, but not after a single conditioning session. Conversely, overshadowing by a taste given after the target was detectable after a single conditioning trial (Experiment 3) and, thus, indicated that retroactive interference involves a process different from that producing proactive interference. A simulation confirmed that the results are consistent with a modified Rescorla and Wagner (1972) interpretation of Revusky's (1971) concurrent interference theory of delay learning.