Pre-exposure Schedule Effects on Generalization of Taste Aversion and Palatability for Thirsty and Not Thirsty Rats

Angulo, Rocío

doi:10.3389/fpsyg.2018.00878

ORIGINAL RESEARCH article

Front. Psychol., 05 June 2018

Sec. Emotion Science

Volume 9 - 2018 | https://doi.org/10.3389/fpsyg.2018.00878

This article is part of the Research Topic Research on Emotion and Learning: Contributions from Latin America View all 35 articles

Pre-exposure Schedule Effects on Generalization of Taste Aversion and Palatability for Thirsty and Not Thirsty Rats

$\r\nRocío Angulo*$ Rocío Angulo^*

Universidad Autónoma de Chile, Santiago, Chile

The study reported four experiments aiming to test the effects of the pre-exposure schedule and water deprivation on the generalization of a conditioned taste aversion in rats, with a particular focus on testing whether or not the concurrent schedule might enhance generalization. In two experiments, non-water-deprived rats received concurrent, intermixed, or blocked exposure to a sweet-acid solution and a salty-acid solution before conditioning of one of these compounds and testing of both flavors. During pre-exposure, the rats consumed a greater amount of the sweet-acid solution than the salty-acid solution (Experiments 1 and 2), consumption of the former increasing during pre-exposure while consumption of the latter decreased (Experiment 1). Furthermore, consumption of the salty-acid solution was lower during concurrent than intermixed or blocked pre-exposure (Experiment 1 and 2) while consumption of the sweet-acid solution was greater during intermixed than concurrent or blocked pre-exposure (Experiment 1). It is discussed whether the pre-exposure schedule might modify stimulus perception beyond the mere enhancement of stimulus differentiation, by, for instance, affecting the palatability of gustatory stimuli. Evidence for enhanced generalization after concurrent pre-exposure was not found for either deprived (Experiments 1, 2, and 3) or non-deprived rats (Experiments 3 and 4), with deprivation leading to a general increase in consumption of both the conditioned and test flavors. This then raised the question of whether or not concurrent pre-exposure to flavors always increases generalization between them. The present study highlights the importance of this issue for various accounts of perceptual learning.

Introduction

A conditioned taste aversion established to one stimulus (e.g., AX) will, to some degree, generalize to another similar stimulus (e.g., BX), depending on the schedule with which the stimuli have been presented previously (e.g., Honey and Hall, 1989; Mackintosh et al., 1991). In particular, generalization between the stimuli appears to be lower after intermixed (i.e., AX, BX, AX, BX…, the stimuli being presented few hours apart) than blocked (i.e., AX, AX,…, BX, BX…) pre-exposure (e.g., Symonds and Hall, 1995, 1997; Mondragón and Hall, 2002). But shorter intervals (a few minutes or seconds) between intermixed presentations of the stimuli (e.g., Bennett and Mackintosh, 1999; Wills and Mackintosh, 1999) or concurrent pre-exposure (e.g., Alonso and Hall, 1999; Rodríguez and Alonso, 2008), seems to increase generalization with respect to the standard intermixed or blocked schedules.

Within the field of perceptual learning, it is widely accepted that generalization decrements after stimulus pre-exposure must rely, at least in part, on changes in the extent to which the stimuli can be differentiated as a consequence of stimulus pre-exposure (see Hall, 1991; Mitchell and Hall, 2014, for a review). To the extent that less generalization would be expected between stimuli that are better differentiated from each other, then such an effect should be more marked when using an intermixed pre-exposure schedule, compared to a schedule in which the stimuli are pre-exposed in a series of blocks. According to such an analysis, it is expected that stimulus differentiation would be poorer after rapid-intermixed or concurrent pre-exposure relative to standard intermixed or blocked pre-exposure. A number of human perceptual learning experiments, however, have found quite the opposite result. In particular, several studies have established that people are more accurate in judging two similar visual stimuli as different, or assigning such stimuli to different artificial categories following concurrent exposure in comparison with either intermixed or blocked schedules (e.g., Mundy et al., 2007, 2009; Angulo and Alonso, 2012, 2013). It has therefore been concluded that at least for humans, the concurrent schedule increases stimulus differentiation to a greater extent than the intermixed or blocked schedules. This apparent discrepancy between the results of studies conducted with human and non-human animals has been explained in several ways.

For example, it has been argued that, even in the case in which concurrent pre-exposure is able to enhance stimulus differentiation for non-human animals in the way described for people, various sources of generalization might still be enhanced by this schedule. According to some authors, concurrent pre-exposure to the stimuli will strengthen the associative excitatory links between them (e.g., Honey et al., 1994; Alonso and Hall, 1999; Rodríguez et al., 2008) and, such links would be able to increase generalization between the stimuli, even in the case in which the subjects are able to discriminate between them. Indeed, it is expected that excitatory links would be established between concurrently presented stimuli only in the case in which such stimuli were discriminable. It has also been suggested that common elements of the stimuli (X, in the example outlined above) could be more strongly conditioned after concurrent than intermixed or blocked pre-exposure (e.g., Bennett and Mackintosh, 1999; Rodríguez and Alonso, 2008). Thus, irrespective of how well animals may differentiate between the stimuli after concurrent pre-exposure, the size of the generalized conditioned response (CR) will be stronger, simply because learning acquired by the common elements would be greater.

Other analyses have focused on the extent to which the instructions given to participants during the pre-exposure phase could play a role in determining the outcome of human perceptual learning studies. To date, human studies assessing the concurrent schedule (e.g., Mundy et al., 2007, 2009; Angulo and Alonso, 2012, 2013) have specifically asked participants to look for differences between the stimuli during pre-exposure, while for other animals, stimuli are assumed to be “merely” pre-exposed (e.g., Hall, 1991, 2001). It has been proposed that, in humans, instructions may activate top-down strategies for processing the stimuli that would not be operating in situations in which mere exposure is given to other animals (e.g., Mitchell and Hall, 2014). According to such analyses, the greatest benefit for humans (although not for other animals) of the concurrent schedule may lie in an interaction of some sort between the pre-exposure schedule and the instructions asking the participant to seek out stimulus differences (see Nelson, 2009; Mitchell and Hall, 2014, for a review of this discussion). A recent study conducted with humans found that, in the absence of instructions asking participants to specifically look for stimulus differences, people were less accurate in judging the pre-exposed stimuli as being same or different, following concurrent pre-exposure (Angulo et al., unpublished). It might therefore be tempting to conclude that the effects of the pre-exposure schedule could vary according to the specific demands of the instructions given during pre-exposure and, if this were indeed the case for humans, one might wonder whether the same may also be true for other animals. Rats and other non-human animals cannot be verbally instructed during pre-exposure. But it could still be the case that the animals’ environment (external or internal) during pre-exposure may dictate the way in which the stimuli are processed, in the same way that verbal instructions are thought to influence stimulus processing in humans.

Experimental procedures employing taste aversion preparations with rats usually begin with a regime of water deprivation that is maintained throughout the experiment (e.g., Symonds and Hall, 1995, 1997; Honey and Hall, 1989; Mackintosh et al., 1991; Rodríguez and Alonso, 2008). Thus, the procedure induces a state of fluid deprivation under which the critical flavored solutions will be presented. In such a situation, rats receive exposure to stimuli with motivational value in a state of fluid depletion. It might therefore be a matter of debate as to what extent the rats are receiving mere exposure to neutral stimuli, as has traditionally been assumed. It might be possible to argue that the state of water deprivation — a part of the animal’s internal environment — might provide some “like-instructions” for processing the stimuli. Accordingly, the aim of the present study was to examine the impact of giving the animals free access to water during the pre-exposure phase (i.e., not inducing a state of water deprivation). In particular, we wanted to explore whether this change in the deprivation state would allow us to observe differences in the extent to which the schedule of pre-exposure has an effect on stimulus generalization, using a standard taste aversion procedure. Because discrepancies between the studies conducted with human and non-human animals arise from the effects of the concurrent pre-exposure schedule, we focused particularly on the effects of this schedule.

Experiment 1

Using a conditioned taste aversion procedure, Experiment 1 tested the effect of concurrent, intermixed, and blocked pre-exposure to two similar flavored compound solutions, AX and BX, on subsequent generalization between them. Whilst this issue has previously been addressed (Rodríguez and Alonso, 2008; Rodríguez et al., 2008), the rats were fluid deprived during pre-exposure (but see Alonso and Hall, 1999 for concurrent and blocked pre-exposure effects when the flavored compounds were presented ad libitum). In this experiment, however, the deprivation regime began after pre-exposure. The experiment included eight groups of rats. Half of these groups received a single taste aversion conditioning trial with the compound AX (paired groups, - P) after pre-exposure, while the other half received presentation of the stimulus AX and the induced illness separately (unpaired groups, - U). Within the paired and unpaired groups, one received a concurrent pre-exposure schedule (e.g., AX–BX, AX–BX, AX–BX, etc., Groups CNC-P and CNC-U), another received the intermixed schedule (e.g., AX, BX, AX, BX, etc., Groups INT-P and INT-U) and the third the blocked schedule (e.g., AX, AX, etc., BX, BX, etc., Groups BLK-P and BLK-U). The remaining two groups received no pre-exposure to AX and BX before the induced illness (Groups CTRL-P and CTRL-U). Subsequently, all the rats received a single consumption test with the AX flavor, and another test with BX. Generalization between these stimuli was then assessed in two ways: by comparing AX and BX consumption in the paired groups; and by comparing consumption of BX in the paired groups with consumption of BX in the unpaired groups. Different consumption levels of AX and BX in the paired groups would indicate discrimination between the two stimuli. Since AX but not BX was aversively conditioned, lower consumption of AX than BX would be expected if rats were indeed able to differentiate between the stimuli. No differences between consumption of AX and BX in the paired groups would, however, be taken to indicate generalization between the stimuli. On the other hand, lower consumption of BX in the paired than in their respective unpaired control groups would also be indicative of generalization. Since no stimulus was conditioned in the unpaired groups, and thus there would be no aversion to be generalized from one stimulus to the other, consumption of BX and AX should be similar in this case, with consumption of BX being of the level that would be expected in the case of “zero generalization.”

Experiment 1 also measured the free consumption of the two flavored solutions employed as stimuli, AX and BX, during concurrent, intermixed, and blocked pre-exposure. The reason for this is that it has recently been suggested (Mitchell and Hall, 2014), though not empirically tested, that rats might be able to differentiate a priori the stimuli usually employed in perceptual learning studies. For the current theoretical accounts of perceptual learning (e.g., McLaren and Mackintosh, 2000; Hall, 2003; Mitchell et al., 2008), it becomes crucial to test this possibility. Such accounts have been developed primarily within the framework of animal research and have provided evidence for mechanisms that are able to explain stimulus differentiation as a consequence of pre-exposure. But, if the stimuli are differentiated a priori, then differentiation must clearly not be a consequence of such proposed mechanisms. As in previous studies (e.g., Mackintosh et al., 1991; Symonds and Hall, 1995), the two flavored compound solutions used here as stimuli shared a common acid flavor (X), and were distinguished by the addition of other distinctive flavors (A and B), which are sugar and salt (counterbalanced in each group, see below). The sweet solution was expected to be more palatable than the salty solution. Thus, if rats were able to differentiate between the stimuli during pre-exposure, and had access to an unrestricted amount of the solution, they should display a preference for the sweet solution relative to the salty one. To the best of our knowledge, pre-exposure schedule effects on stimulus discrimination have not yet been tested in this simple and more ecological way.

Method

Subjects and Apparatus

Subjects were 64 experimentally naïve male Wistar rats with a mean weight of 350 g (range 346–355 g) at the beginning of the experiment. Rats were individually housed in cages with food and water ad libitum in a room with constant temperature (24°C) and humidity (50%). The room was artificially illuminated under a 12 h-dark/light cycle with the light period beginning at 8:00 a.m. All experimental sessions were conducted with the animals in their home cages.

The stimuli used were two compound solutions, AX and BX, consisting of diluted plain water presented at room temperature through 50-ml plastic tubes, fitted with a metal spout. Stimulus X was always a solution containing hydrochloric acid (HCl) at 1%, whilst Stimulus A and B were (counterbalanced) a solution of 1% salt and 10% sugar, respectively, for half the rats in each group, and the reverse for the remaining half. The unconditioned stimulus (US) was an intraperitoneal injection of 0.15 M lithium chloride (LiCl) administered at 10 ml/Kg of body weight.

Procedure

The Animal Welfare Ethics Committee of the Universidad del Pais Vasco (Experiments 1 and 2) and the Universidad Autónoma de Chile (Experiments 3 and 4) approved the procedures employed in the current experiments. The rats were randomly assigned to eight equal groups (n = 8): CNC-P, INT-P, BLK-P, CTRL-P, CNC-U, INT-U, BLK-U, and CTRL-U.

Pre-exposure

All the rats, except those belonging to Groups CTRL-P and CTRL-U, received four AX and four BX presentations during pre-exposure. This phase lasted for 4 days, with the stimuli being presented during two daily 30-min drinking sessions (at 11:00 a.m. and 5:00 p.m.). These drinking sessions began by removing the bottles containing ad libitum plain water, and replacing these with two tubes, each containing 50 ml of flavored solution, one located on the right-hand side and the other on the left-hand side of the rats’ home-cages (30 cm apart). The sessions finished with the replacement of the two tubes with the bottle containing water ad libitum. For the CNC-P and CNC-U groups, in all the pre-exposure sessions one of the bottles contained the salty compound and the other the sweet compound. Half of the rats in these two groups first received the salty flavor on the right and the sweet flavor on the left (with the remaining rats receiving the reverse arrangement), the position of the solution being switched from trial to trial. For the INT-P and INT-U groups, the two tubes always contained the same solution, but it was switched from trial to trial. Half of the rats received the salty flavor on all of the morning sessions and the sweet flavor on the afternoon sessions, while for the remaining rats this arrangement was reversed. The BLK-P and BLK-U groups also received the same solution in both tubes. However, in these groups, half of the rats received the salty flavor on the first 2 days (on both the morning and afternoon sessions) and the sweet solution on the latter 2 days, while the other half received the reverse arrangement. Finally, the CTRL-P and CTRL-U groups always received plain water in both tubes. For half the rats in each group the salty solution was flavor AX and the sweet solution the flavor BX, while the reverse was true for the remaining rats. One hour after the last pre-exposure drinking session, the bottles containing ad libitum water were removed from the home cages. On the following 4 days, all the rats only had access to fluids during the 30 min morning and afternoon drinking sessions, in a single bottle placed in the location otherwise occupied by the standard water bottle.

Conditioning

The day after the last pre-exposure trial, all the rats received a single intraperitoneal injection of LiCl, half of them in the morning session and the other half in the afternoon session. For the CNC-P, INT-P, BLK-P, CTRL-P, and CNC-P groups, the injection was immediately preceded by the presentation of 10 ml of the AX solution in a single bottle, with the rats receiving 10 ml of plain water on the other session. Half of the rats from Groups CNC-U, INT-U, BLK-U, and CTRL-U received 10 ml of the AX solution in the morning, and 10 ml of plain water followed by the injection in the afternoon, whilst the remaining rats received 10 ml of water followed by the injection in the morning and 10 ml of the AX solution in the afternoon. The day following conditioning, all the rats received free access to plain water in the morning and afternoon drinking sessions.

Test

Over the next 2 days, all the rats had access to free consumption of the AX or BX solutions in the morning drinking session and water in the afternoon. Half of the rats from each group received the AX solution on the first day and the BX solution on the second, with the remaining rats receiving BX first and then AX. For all the subjects, the fluid was presented in a single bottle.

The amount of fluid consumed during the drinking sessions was calculated by weighing the drinking tubes before and after consumption and converting the difference to ml. An analysis of variance (ANOVA) was then conducted on the consumption values, adopting a statistical significance criterion of p < 0.05. Subsequent pair-wise comparisons between groups were conducted with the Tukey’s Honest Significant Difference (HSD) test.

Results

The mean consumption of the salty and sweet solutions during pre-exposure trials with the three schedules (Concurrent, Intermixed, and Blocked) is shown in Figure 1. Consumption of the sweet solution was always greater than consumption of the salty solution, with this consumption increasing in the former case and decreasing in the latter across the blocks of pre-exposure trials. Furthermore, consumption of the sweet solution was greater with the intermixed pre-exposure schedule than with the others, while consumption of the salty solution was lower with the concurrent schedule than the other schedules.

FIGURE 1

FIGURE 1. Consumption (ml) of the salty and sweet solutions for the concurrent, intermixed, and blocked pre-exposure conditions across blocks of pre-exposure trials in Experiment 1.

A 3 × 2 × 4 ANOVA with Schedule, Solution, and Trial as factors found a significant main effect of the three factors, Schedule, F(2,45) = 10.99, p < 0.001; Solution, F(1,45) = 4.15, p < 0.001; and Trial, F(3,135) = 4.89, p = 0.003. The Schedule × Solution, F(2,45) = 3.62, p = 0.035, and Solution × Trial, F(3,135) = 13.29, p < 0.001, interactions were also significant. A subsequent analysis of simple effects for these interactions confirmed the observations described above. Greater consumption of the sweet solution than the salty solution was found with all three pre-exposure schedules, Fs(1,15) ≥ 95.89, ps ≤ 0.001, with pre-exposure conditions differing in both cases, Fs(2,45) = 10.25, p < 0.001. Consumption of the sweet solution was greater with the intermixed than with the concurrent, p = 0.003, or the blocked schedule, p = 0.007, while consumption of the salty solution was lower with the concurrent than with the intermixed, p < 0.001, or blocked, p = 0.005, schedule. Consumption of the sweet solution was greater than the salty solution in all four blocks of pre-exposure trials, Fs(1,47) ≥ 44.47, p ≤ 0.001, with the former increasing across blocks of trials, F(3,141) = 11.36, p < 0.010, while the latter decreased, F(3,141) = 3.95, p = 0.010.

On the conditioning trial, consumption of AX was 8.04 (SEM ± 1.36), 9.66 (SEM ± 1.29), 9.86 (SEM ± 1.53), and 8.77 (SEM ± 1.43) for the Concurrent, Intermixed, Blocked, and Control conditions, respectively. An ANOVA conducted with these consumption data found no significant effect of Schedule, F(3,60) = 0.35, p = 0.783.

The mean consumption of AX and BX for the paired and unpaired groups during testing can be seen in Figure 2. Irrespective of the pre-exposure schedule, unpaired groups consumed a similar amount of AX and BX, but consumption of AX was always lower than consumption of BX for the paired groups. In this latter case, consumption of AX was similar in all groups, while consumption of BX appeared to be greater for the CNC-P and BLK-P groups than for the INT-P and CTRL-P groups. Furthermore, consumption of AX was, in general, lower for the paired than for the unpaired groups. However, consumption of BX was observed to be somewhat lower for the INT-P and CTRL-P groups than for the INT-U and CTRL-U groups, respectively, and very similar for the CNC-P and CNC-U, and BLK-P and BLK-U groups. A 4 × 2 × 2 ANOVA conducted with Schedule, Stimulus, and Conditioning as factors supported all the above observations except for the latter. The main effect of Conditioning, F(1,56) = 21.52, p < 0.002, and Stimulus, F(1,56) = 11.28, p = 0.001, as well as the interaction Stimulus × Conditioning, F(1,56) = 21.211, p < 0.001, were significant (remaining Fs ≤ 0.89). A subsequent analysis of simple effects for the Stimulus × Conditioning interaction found lower consumption of AX than BX in the paired groups, F(1,31) = 41.28, p < 0.001, but not in the unpaired groups, F(1,31) = 0.616, p = 0.439. Furthermore, consumption of AX was lower in the paired than in the unpaired groups, F(1,62) = 45.57, p < 0.001, while consumption of BX did not differ reliably in the paired and unpaired groups, F(1,62) = 0.470, p = 0.496. Previous studies assessing the effect of the concurrent schedule on generalization have inferred its magnitude only from the consumption of BX (e.g., Alonso and Hall, 1999; Rodríguez and Alonso, 2008). Thus, although the Schedule × Stimulus interaction failed to reach significance, some planned comparisons were conducted in order to explore the differences in consumption observed for the paired groups. These analyses indicated that while the paired groups did not differ in their consumption of AX, F(3,28) = 0.26, p = 0.84, they differed in their consumption of BX, F(3,28) = 4.24, p = 0.014. In particular, consumption of BX was greater in Group BLK-P than in Group CTRL-P (p = 0.026).

FIGURE 2

FIGURE 2. Consumption (mean ± SEM) of the AX and BX solutions during the test trials for the 8 groups of Experiment 1.

Discussion

In Experiment 1, rats consumed a greater amount of the sweet-acid solution than the salty-acid solution from the first pre-exposure trial. This initial discriminative response to the solutions seems to indicate that the rats were able to differentiate between them a priori, the sweet solution being more palatable than the salty solution. The consumption of the salty solution decreased in general during pre-exposure while consumption of the sweet solution increased. Thus, at least when rats were not water-deprived during pre-exposure, consumption of the less palatable solution might decrease unconditionally, while consumption of the more palatable solution increases. Experiment 1 also found a clear effect of the schedule on the pattern of consumption elicited by the salty and sweet solutions during pre-exposure. Rats consumed less of the salty solution during concurrent than intermixed or blocked pre-exposure. However, the sweet solution was more readily consumed during intermixed than concurrent or blocked pre-exposure. If we accept that the stimuli were readily discriminated a priori, it is possible that the schedule affected consumption of the most and least palatable stimuli not by increasing stimulus differentiation, but rather through some other, as yet unspecified, effect of the schedule. This issue will be taken up again in the General Discussion.

On the test, all the rats from the paired groups drank less of the aversively conditioned flavor AX than the BX flavor. Again this discriminative response would be possible only if the stimuli were differentiated. Rats from the unpaired groups did not differ in their consumption of AX and BX. Given that no stimulus was conditioned in these groups, there would be no reasons to expect any difference in consumption of AX and BX even when they are discriminable (note that half of the rats that received the AX compounds were given the sweet solution and the other half received the salty solution). Consumption of AX was lower in the paired than in the unpaired groups, indicating that an aversion to AX was successfully established in these cases. And importantly, there were no significant differences between the paired and unpaired groups in terms of their consumption of BX. Since consumption of BX in the unpaired groups would be free from any generalization of conditioned aversion, this last result may indicate that the paired groups did not generalize the conditioned aversion from AX to the other stimulus, BX. In brief, none of the results obtained on the test appear to indicate generalization between AX and BX in the paired groups, and we can thus rule out the possibility that generalization was differentially affected by the pre-exposure schedules. Nevertheless, in the paired groups, consumption of BX was observed to be greater after blocked pre-exposure to the stimuli relative to the control group that did not receive prior pre-exposure to the stimuli. Taking the consumption of BX alone, it might be possible to conclude that the blocked but not the intermixed and concurrent pre-exposure schedules reduced generalization between the stimuli (relative to the control group not receiving pre-exposure to the stimuli). Such a conclusion would be incompatible with the results produced by previous studies (Alonso and Hall, 1999; Rodríguez and Alonso, 2008; Rodríguez et al., 2008), in which rats drank less BX after concurrent than intermixed or blocked pre-exposure, concluding that generalization was greater in the former than in the latter cases. However, the results of Experiment 1 appear to suggest that consumption of BX might not be a net indicator of generalization between the stimuli. Given that there is no sign of generalization between the stimuli, and taking into account that pre-exposure schedules might directly affect consumption, it seems likely that differences in consumption of BX between the groups could reflect some effect of the schedules on consumption rather than on stimulus generalization.

Presumably, the discrepancy between our findings and those of previous studies regarding the consumption of BX might be related to the state of fluid deprivation during pre-exposure. However, it should be noted that this and the previous studies also differed in terms of the amount of solution that was made available during pre-exposure. In the experiments conducted previously, rats received only a small and fixed amount of the solutions during pre-exposure (10 ml), and these were usually fully consumed given that the rats were water deprived (but see Alonso and Hall, 1999, Experiment 1a). But in Experiment 1 of the present study, rats had free access to the flavored solutions. Therefore, the main aim of Experiment 2 was to test whether the principal findings of Experiment 1 could be replicated if rats received a fixed amount of the solutions during pre-exposure, as in previous experiments reported in the literature.

Experiment 2

The methods used in Experiment 1 and those employed in previous studies (e.g., Rodríguez and Alonso, 2008; Rodríguez et al., 2008, but see Alonso and Hall, 1999), differed in two important ways. Firstly, in Experiment 1 (but not in previous studies), the rats were not in a state of water deprivation during pre-exposure (but see Alonso and Hall, 1999). And secondly, in Experiment 1 (but not in the previous studies), rats were allowed free access to the solutions during pre-exposure instead of a limited, fixed amount of the flavor. In Experiment 2, rats in the paired groups also received a limited amount of the solution during pre-exposure in order to see if such a variable could play a role in our failure to observe the expected effect of enhanced generalization after concurrent pre-exposure.