Conditioned taste aversion and traditional learning

Summary The generality of the laws of learning has been questioned on the basis of the results of conditioned taste aversion (CTA) experiments. The results of CTA experiments and of traditional learning (TL) experiments are compared in order to evaluate whether CTA and TL indeed follow different rules. It is shown that (a) most of the observed differences between CTA and TL experiments may simply reflect peculiarities of the CSs and USs used in CTA, (b) that there are conflicting phenomena in both CTA and TL experiments which require further analysis, and (c) that the majority of learning phenomena seen in CTA experiments also occur in TL experiments. It is therefore suggested that the laws that hold for TL also apply to CTA.     

Context-dependent latent inhibition in taste aversion learning

The effects of taste stimulus preexposure, either in the presence or in the absence of a specific contextual cue consisting of a specific noise-producing bottle, upon the conditioning and testing of conditioned taste aversions to the taste (saccharin) plus context (noisy-bottle) compound stimulus were investigated. Four groups of rats were given preexposure trials (latent inhibition) to either: (1) novel saccharin in novel noisy bottles, (2) novel saccharin in familiar silent bottles, (3) familiar water in novel noisy bottles, (4) familiar water in familiar silent bottles, in six trials. During conditioning, saccharin was presented in the noisy bottles followed by lithium chloride for all the groups. At testing, saccharin was presented in the noisy bottles for both one-bottle and two-bottle tests of aversion. It was indicated that the conditioning decrement produced after both saccharin and noisy bottle preexposure was overwhelmingly greater than any produced after preexposure to the elements. These results, discussed in relation to current theories of latent inhibition and perceptual learning, further underline the overwhelming significance of exteroceptive contextual elements in conditioned taste aversions.     

Cortical substrates of taste aversion learning: dorsal prepiriform (insular) lesions disrupt taste aversion learning

The functional relation between restricted damage to ventral primary somatosensory neocortex and the ability of rats to acquire conditioned taste aversion (CTA( was examined by a combination of behavioral and neurohistological techniques. Lesions confined exclusively to the established gustatory neocortex (GN) did not disrupt CTA acquisition, nor did lesions confined to suprarhinal cortical areas ventral to the GN. Lesions that encroached on dorsal prepiriform and insular cortices produced CTA acquisition deficits and damaged a large proportion of efferent projections to the prefrontal and precentral neocortex. In a second experiment, lesions of dorsal prepiriform and insular cortices did not modify taste preference-aversion threshold to any of the four taste modalities. It is concluded tha ventral somatosensory neocortical fields, including the established GN, do not mediate CTA acquisition and that rhinal cortices ventral and posterior to the GN are preferentially involved in associative learning for tastes and illness.     

Within-compound associations of taste and temperature

In two experiments, rats were given exposures to two solutions of different tastes (sucrose or sodium chloride) at two different temperatures (warm or cold). In Experiment 1, the rats were then given either one of the tastes at room temperature followed by LiCl injections, or distilled water at one of the temperatures followed by LiCl injections. Rats poisoned after drinking a taste were then given a choice between distilled water at the two temperatures, while those poisoned after drinking distilled water at a temperature were given a choice between the two flavors at room temperature. Rats drank less of the solution that contained the stimulus previously paired with the poisoned cue, demonstrating within-compound associations of tastes and temperatures. Experiment 2 found that tastes were better conditioned in a taste aversion procedure when the taste-temperature compound was the same during conditioning as during preexposure. This result is interpreted as evidence against a view of within-compound learning that treats the compound as a unitary stimulus.     

Dorsomedial pallium lesions impair taste aversion learning in goldfish

The present work shows that the dorsomedial telencephalic pallium of teleost fish, proposed as homologous to the amygdala of mammals, is involved in taste aversion learning (TAL). To analyze the behavioral properties of TAL in goldfish, in Experiment 1, we used a delayed procedure similar to that employed with mammals, which consists of the presentation of two flavors on different days, one followed by lithium chloride and the other by saline, both after a 10-min delay. The results showed that goldfish developed a strong aversion to the gustatory stimulus followed by visceral discomfort and that, as in mammals, this learning was rapidly acquired, highly flexible and maintained for a long time. Experiment 2 showed that dorsomedial pallium lesions and the ablation of the telencephalic lobes impaired the acquisition of taste aversion in goldfish, whereas damage to the dorsolateral pallium (hippocampus homologue) or cerebellar corpus did not produce significant changes in this learning. Experiment 3 showed that these TAL deficits were not due to a lesion-related disruption of taste discrimination; goldfish with telencephalon ablation were able to learn to distinguish between the two tested flavors in a differential conditioning procedure. These functional data demonstrate that the dorsomedial pallium in teleosts is, like the amygdala, an essential component of the telencephalon-dependent taste aversion memory system and provide further support concerning the homology between both structures.     

Context specificity of latent inhibition in taste aversion learning

In three experiments rats were given injections of LiCl after consuming distinctively flavoured water. The rats developed an aversion to the flavour and in all experiments the magnitude of the aversion was found to be reduced in subjects that had received pre-exposure to the flavour without aversive consequences. Experiment 1 demonstrated this pre-exposure effect to be a case of latent inhibition. The remaining experiments investigated the effects of pre-exposing the flavour in a context different from that used for conditioning. It was found (Experiment 2) diat latent inhibition transferred perfectly when the context change consisted of a move from one home cage to another. Context specificity of latent inhibition was found (Experiment 3) only when the subjects were given daily sessions in die experimental contexts, these being cages different from the home cage.     

Impaired conditioned taste aversion learning in spinophilin knockout mice

Plasticity in dendritic spines may underlie learning and memory. Spinophilin, a protein enriched in dendritic spines, has the properties of a scaffolding protein and is believed to regulate actin cytoskeletal dynamics affecting dendritic spine morphology. It also binds protein phosphatase-1 (PP-1), an enzyme that regulates dendritic spine physiology. In this study, we tested the role of spinophilin in conditioned taste aversion learning (CTA) using transgenic spinophilin knockout mice. CTA is a form of associative learning in which an animal rejects a food that has been paired previously with a toxic effect (e.g., a sucrose solution paired with a malaise-inducing injection of lithium chloride). Acquisition and extinction of CTA was tested in spinophilin knockout and wild-type mice using taste solutions (sucrose or sodium chloride) or flavors (Kool-Aid) paired with moderate or high doses of LiCl (0.15 M, 20 or 40 mL/kg). When sucrose or NaCl solutions were paired with a moderate dose of LiCl, spinophilin knockout mice were unable to learn a CTA. At the higher dose, knockout mice acquired a CTA but extinguished more rapidly than wild-type mice. A more salient flavor stimulus (taste plus odor) revealed similar CTA learning at both doses of LiCl in both knockouts and wild types. Sensory processing in the knockouts appeared normal because knockout mice and wild-type mice expressed identical unconditioned taste preferences in two-bottle tests, and identical lying-on-belly responses to acute LiCl. We conclude that spinophilin is a candidate molecule required for normal CTA learning.     

Discriminative taste aversion learning: a learning task for older chickens

The study of learning and memory using the chicken model has relied on three learning paradigms, passive avoidance learning, imprinting and the pebble floor task. Passive avoidance learning and imprinting have been used predominantly in very young chickens and cannot be used to access learning and memory in older chickens. We have established a new behavioural learning paradigm, Discriminative Taste Aversion Learning (DTAL), that can be used with both young and older animals. The task requires chickens to discriminate between food crumbs dyed either red or yellow with one colour being associated with the aversive tasting substance, methylanthranilate. Learning can be tested at various times after the training session by presenting chickens with the coloured food crumbs without an aversive taste. Both chickens tested at 5 and 15 days post-hatch learned to avoid the aversive crumbs. Furthermore, the protein synthesis inhibitor anisomycin (30 mM; 10 microl per hemisphere) injected into the intermediate medial hyperstriatum ventrale 15 min pre-training or 45 min post-training blocked long-term memory for the DTAL task when tested 24 h later. Memory for the task was unaffected by anisomycin injection 120 min post-training or in control animals injected with saline at similar times. The timing of the cellular processes of protein synthesis needed for consolidation of the DTAL appears to be similar to those described for the other behavioural paradigms in young chickens.     

Conditioned taste aversion in humans using motion-induced sickness as the US

The purpose of the experiment was to demonstrate conditioned taste aversion (CTA) and latent inhibition (LI) of CTA in humans using rotation-induced motion sickness as the unconditioned stimulus. To accomplish this, flavour familiarity (familiar vs unfamiliar) and rotation (rotation vs no rotation) were manipulated in a 2 X 2 factorial design. Subjects consumed either a familiarly flavoured carbonated beverage or a novel one after which half of each group was rotated or not rotated. Two hours later the subjects were re-presented with the flavoured drink that they had previously drunk. The groups receiving rotation consumed less of the drink that the non-rotated groups, thus demonstrating CTA. The rotated group pre-exposed to the novel flavoured drink consumed less than the rotated group pre-exposed to the familiar drink, thus demonstrating LI. The effectiveness of the rotation procedure in producing motion sickness was confirmed by self-reports of general feelings and by symptom rating scales. In addition, it was found that, at the time of consuming the test drink the rotation groups' motion-sickness symptom scores were reduced to the level of the nonrotated groups. Applications of these data to the prophylactic treatment of chemotherapy-induced food aversions were discussed.     

Conditioned taste aversion learning in leptin-receptor-deficient db/db mice

The db/db mouse has defective leptin receptors. The defects lead to impairments of leptin regulation of food intake and body weight, and result in the expression of diabetic symptoms such as hyperinsulinemia, hyperglicemia, and extreme obesity. Recent studies have proposed that leptin may also affect memory and learning processes. To examine this possibility, we compared the ability of leptin-receptor-deficient db/db mice and their normal lean litter mates to form and extinguish a conditioned taste aversion (CTA) for saccharin. We used a short-term (10 s) lick test and a long-term (48 h) two bottle preference test for measurement of consumption of test solutions. On the first day after conditioning to avoid saccharin, the db/db mice showed preference scores for saccharin as low, and aversion thresholds for sucrose lower than that of the lean mice. During the extinction test trials beginning from the second up to the 30th day after conditioning, numbers of licks and preference scores for aversive saccharin and sucrose appeared to be larger, and recovered faster to the control levels in db/db mice. These results indicate that db/db mice with leptin-receptor-deficiency may show equal capacity to form CTAs for saccharin, greater generalization from saccharin to sucrose, and a faster rate of extinction. This suggests that disruption of leptin signalling does not inhibit acquisition of CTA learning, but impairs its extinction. This differential contribution of the leptin system on CTA processes may be due to differential distribution of leptin receptors in the CTA-related brain areas.     

Capsaicin-sensitive afferent vagal fibers are involved in concurrent taste aversion learning

Taste aversion learning (TAL) is a type of learning characterized by rejection of a gustatory/flavor stimulus as a consequence of its pairing with visceral discomfort and malaise. TAL can be established in the laboratory by two different behavioral procedures, concurrent or sequential. Neural mechanisms of these learning modalities remain to be elucidated, but several studies have discussed the implication of various anatomical structures, including the vagus nerve. The aim of this study was to examine the role of capsaicin-sensitive vagal afferent fibers in concurrent (Experiment 1) and sequential (Experiment 2) TAL in Wistar rats. Results showed that perivagal administration of capsaicin (1mg of capsaicin dissolved in 1ml of vehicle (10% Tween 80 in oil)) blocked acquisition of concurrent but not sequential TAL. These data support the hypothesis of two different modalities of TAL mediated by distinct neurobiological systems, with vagal nerve participation only being essential in concurrent TAL.     

Inferior olive lesions impair concurrent taste aversion learning in rats.

Taste aversion learning can be established according to two different procedures, concurrent and sequential. For the concurrent task, two different taste stimuli are offered at the same time, one associated with simultaneous intragastric administration of an aversive stimulus and the other associated with physiological saline. This discrimination is learned by sham-lesioned control animals and by animals with lesions in the cerebellar cortex but not by rats lesioned in the inferior olive. At the same time, animals with lesions in the inferior olive and sham-lesioned animals achieve sequential learning when the gustatory stimuli are offered individually during each daily session. The results obtained show that electrolytic lesions in the inferior olive impair acquisition of concurrent learning and are analyzed in terms of an anatomical system consisting of the vagus nerve, inferior olive, and cerebellum, which differentiates between the two modalities of taste aversion learning, concurrent and sequential.     

Reversal of long-delay conditioned taste aversion learning in rats by sex hormone manipulation

Conditioned taste aversion (CTA) learning is an adaptive, robust, well-established learning and memory paradigm. Strong taste, aversions develop to the conditioned stimulus (CS=saccharin) despite long delays between exposure to the CS and unconditioned stimulus (US=LiCl). Rats display a sexually dimorphic pattern of long-delay CTA learning (Foy et al., 1996). The present study examines whether this sex difference is a result of activational or organizational hormone action, because here we implanted gonadectomized rats with their normal hormone replacements, or with opposing hormones prior to testing in a 4-hr delayed CTA learning task. We found that gonadally intact male rats displayed a more robust CTA response than intact female rats. Gonadectomy essentially eliminated this sex difference; gonadectomized males and gonadectomized females displayed similar CTA responses. In gonadectomized rats, when their normal sex hormones were replaced with implanted hormone pellets, the sex difference in CTA learning was reinstated. In contrast, when gonadectomized rats were implanted with opposing hormones, the sex difference was reversed. Gonadectomized female rats implanted with 5α-DHT pellets (metabolite of testosterone) displayed a stronger CTA response compared to gonadectomized males implanted with 17β-estradiol pellets. Regardless of the original developmental hormonal environment, our study suggests that an activational manipulation of circulating hormones serves to significantly influence long-delay CTA learning in rats.     

Effects of a flavor-placement reversal test after different modalities of taste aversion learning

Taste aversion learning is induced through two different behavioral procedures: a short-term or concurrent (two-daily flavors) and a long-term or sequential (one-daily flavor) procedure. For the concurrent group of animals, two gustatory/olfactory stimuli are presented separately but at the same time on a daily basis. One is paired with simultaneous intragastric administration of hypertonic NaCl and the other with physiological saline. For the sequential group, the two stimuli are presented on alternate days, one of them followed by intragastric injection of the aversive stimulus and the other by saline, both after a delay of 15 min. The two groups learned the task, but when they were subjected to a flavor-placement reversal test only the sequential group was successful in achieving it. In a second experiment, three groups of animals had to learn concurrent or sequential discrimination tasks (with either simultaneous or delayed administration of the visceral stimulus) using only spatial/proprioceptive cues. The data show that none of the groups learned them under these conditions. The results are discussed in terms of the different modalities of learning. Short-term and long-term taste aversion learning are different in the anatomical structures involved, the number of trials required for acquisition and, as shown in this paper, flexibility.     

Ontogenetic changes in the modulation of taste aversion learning by home environmental cues in rats

Eight experiments using 611 rats as subjects were conducted to define and analyze an age-related phenomenon of conditioned taste aversion. When consumption of sucrose solution was followed by LiCl-induced illness in the animals' home, acquisition of the aversion to sucrose solution was retarded in preweanling (18-day-old) rats. This effect was not found in adults or in slightly older (21-day-old) rats. Place of testing had no effect in the younger two age-groups, but in adults manifestation of the acquired aversion was retarded when they were tested in the home. There was no interaction between place of conditioning and testing for any age. The locus of the environmental influence on conditioning in preweanling rats was found to be the place of tasting rather than place of illness, retention interval, or testing. Also, the effect was found to be invariant under minor variations in familiarization of the animal with the non-home environment. The principle emerging from these data and others is that the home environment can have a significant influence on learning and conditioning in the immature rat.