Michelle Symonds

PERCEPTUAL LEARNING IN FLAVOR AVERSION CONDITIONING : ROLES OF STIMULUS COMPARISON AND LATENT INHIBITION OF COMMON STIMULUS ELEMENTS

Abstract In three experiments, rats received training in which an aversion was established to one flavor and the extent to which this aversion generalized to a second flavor was tested. Experiment 1 showed that nonreinforced preexposure to both flavors resulted in reduced generalization between them. Experiments 2 and 3 demonstrated that this reduction in generalization required the two flavors to be presented on alternate trials during preexposure. Subjects given preexposure consisting of a block of trials with one flavor followed by a block of trials with the other showed the same degree of generalization as subjects given no preexposure. The two schedules of stimulus presentation were equated in the total amount of exposure given to each stimulus, making it unlikely that differences in latent inhibition could be responsible for the difference seen on the test. It is suggested that the opportunity for stimulus comparison offered by the alternating schedule might be important in a process of perceptual learning that is responsible for the reduced generalization.

Interval between preexposure and test determines the magnitude of latent inhibition: Implications for an interference account

The effect of a retention interval on latent inhibition was studied in three experiments by using rats and the conditioned taste-aversion procedure. In Experiment 1, we demonstrated an apparent loss of latent inhibition (i.e., a strengthening of the aversion) in preexposed subjects that experienced a retention interval of 12 days between conditioning and the test. In Experiment 2, we found no effect of this retention interval on the habituation of neophobia produced by the phase of exposure to the flavor. In Experiment 3, we showed that interposing a retention interval between preexposure and conditioning produced effects exactly comparable to those seen in Experiment 1. The implications of these results for rival theories of latent inhibition, as an acquisition deficit or as a case of interference at retrieval, are discussed.

Lithium-induced context aversion in rats as a model of anticipatory nausea in humans

In three experiments, rats received injections of lithium chloride (LiCl) before being exposed to a distinctive context. In a subsequent test, rats given access to sucrose solution in this context consumed less than control subjects given sucrose in another context that had been paired with a saline injection (Experiment 1), or was quite novel (Experiment 2). Experiment 3 demonstrated that a context that had been associated with LiCl would serve to block the acquisition of a conditioned flavor aversion when it was presented immediately after the injection on a flavor-LiCl trial. These results show that a procedure in which rats experience the adverse effects of a lithium injection in the presence of contextual cues is effective in endowing those cues with aversive properties. It is argued that the context evokes a state of conditioned nausea, and the parallel with the clinical phenomenon of anticipatory nausea and vomiting (ANV) in human patients is outlined.

The US-Preexposure Effect in Lithium-Induced Flavor-Aversion Conditioning Is a Consequence of Blocking by Injection Cues

In 2 experiments, rats received flavor-aversion conditioning in which the unconditioned stimulus (US) was an orally consumed solution of lithium chloride (LiCl). The resulting aversion was not attenuated by giving preexposure to injections of LiCl, although such preexposure did attenuate aversions established using injected LiCl as the US (Experiment 1). This outcome suggests that blocking by injection-related cues is responsible for the US-preexposure effect observed in this situation. Experiment 2 confirmed this interpretation by showing that presenting such cues (by giving an injection of saline) at the time that the LiCl was drunk resulted in an attenuation of conditioning in animals preexposed to injections of LiCl. The US-preexposure effect obtained in these experiments can be explained solely in terms of blocking by injection cues, although other mechanisms may contribute to the effect seen in other flavor-aversion paradigms.

Overshadowing and latent inhibition of context aversion conditioning in the rat

A review is presented of experimental studies, using rats as the subjects, that were designed to establish an animal model of the clinical phenomenon of anticipatory nausea. Experiments 1 and 2 demonstrated that pairing a distinctive context with an illness-inducing injection of lithium chloride endowed the context with new properties, consistent with the proposal that classical conditioning had established an association between the context as the conditioned stimulus and nausea as the unconditioned stimulus. The conditioned response to the context constitutes a form of anticipatory nausea. Experiment 3 examined overshadowing, showing that the presence of a novel salient cue (a flavour) during context conditioning reduced the magnitude of the aversion conditioned to the context. Experiments 4-7 examined the effects of giving exposure to the context prior to conditioning. They demonstrated a latent inhibition effect, that is, a reduction in the magnitude of the aversion in pre-exposed animals. It is suggested that these ways of modulating conditioned aversions could form the basis of interventions for use in the chemotherapy clinic. Anticipatory nausea is assumed to be a consequence of the formation of an association between the cues that constitute the clinic and the drug-induced nausea experienced in their presence. By restricting the development of this association, latent inhibition and overshadowing procedures should be effective in alleviating the problem of anticipatory nausea.

Overshadowing not potentiation of illness-based contextual conditioning by a novel taste

In four experiments using rat subjects, we investigated the effects of presenting a novel flavor cue at the time of pairing an environmental context with illness. In each experiment, the subjects were allowed to spend time in a distinctive cage before receiving an injection of LiCl. For some, plain water was made available on these conditioning trials; for others, a novel taste (HCl) was presented. We measured the strength of the context aversion by assessing the ability of the contextual cues to block the acquisition of an aversion to sucrose in a further phase of training. We found that initial training with HCl present made the context less effective as a blocking cue and concluded that the HCl had overshadowed learning about the context. We suggest that this blocking procedure provides a more accurate assessment of contextual aversion than does the consumption test that has more usually been used and that taste-context overshadowing may be a more robust phenomenon than has thus far been thought.

Postinjection suppression of drinking is modified by the presence of conditioned contextual cues: Implications for both anticipatory and posttreatment nausea in humans

In three experiments, we set out to determine whether the response of rats to an injection of LiCl would be modified by the presence of an environmental context that had previously been paired with LiCl. Experiment 1 confirmed that one feature of the malaise produced by LiCl is a reduced tendency to consume an otherwise palatable flavor. Experiment 2 showed that the size of this response was enhanced if it was measured in the presence of a conditioned context. In Experiment 3, we investigated the possibility that the postinjection response could be modified by an overshadowing treatment given during the conditioning phase. The significance of these findings for the understanding of chemotherapyinduced nausea in the clinical population is discussed.

Perceptual Learning With a Sodium Depletion Procedure

Rats in a state of salt need prefer a flavor that has previously been paired with saline (Experiment 1). In Experiments 2 and 3, rats exposed to 2 saline concentrations, presented either concurrently or on separate trials, and each paired with a different flavor, showed a preference for the flavor that had been associated with the stronger saline. This effect was substantial, however, only in those rats that had experienced the concurrent exposure schedule. This effect cannot be attributed to a difference in the strength of within-compound associations produced by the 2 preexposure schedules (Experiment 4). It is suggested that concurrent preexposure can engage a learning process that enhances the discriminability of the preexposed stimuli.

Contextual conditioning with an illness US is attenuated by the antiemetic ondansetron

In two experiments, rats received an injection of LiCl before being placed in a distinctive context. The formation of an aversion to the context was evident in the rats’ unwillingness, on a subsequent test trial, to consume a (normally palatable) sucrose solution in that context. Experiment 1 demonstrated that the size of the context aversion was reduced in rats that received an injection of the antiemetic ondansetron, a selective 5-HT3 antagonist, immediately prior to the injection of LiCl. Experiment 2 confirmed this effect and showed it to occur also in rats that received an injection of ondansetron prior to the test trial (thus ruling out an explanation of Experiment 1 in terms of state-dependency of conditioning). The implications of these findings for the development of an animal learning model of the conditioned side effects experienced by patients undergoing cancer chemotherapy are considered.

Context specificity of sensory preconditioning: Implications for processes of within-event learning

A variant of the taste aversion procedure for sensory preconditioning, as used by Rescorla and Cunningham (1978), was employed in a study of the context dependency of within-event learning. In two experiments, rats received Phase 1 exposure to a simultaneous flavor compound, AX; flavor X was paired with illness during Phase 2, and any tendency for the resulting aversion to be elicited by A was measured. It was found that subjects were less likely to shun flavor A as a consequence of this training if the Phase 1 and test episodes were conducted in distinctively different contexts. This effect was evident both when the change of context occurred just before the test with flavor A (Experiment 1) and when it occurred before Phase 2 (Experiment 2). These results were taken to imply that, as is often found with serial associations, the retrieval of within-event associations is subject to contextual control. The implications of these findings for the interpretation of perceptual learning effects are discussed.