Contextual control of running-based taste aversion in rats.
In poison-based taste-aversion learning, numerous research efforts have documented that background contexts can control the amount of taste intake (e.g., Boakes, Westbrook, Elliott, & Swinbourne, 1997; Bonardi, Honey, & Hall, 1990; Loy, Alvarez, Rey, & Lopez, 1993; Nakajima, Kobayashi, & Imada, 1995; Puente, Cannon, Best, & Carrell, 1988; Sjoden & Archer, 1989; Skinner, Martin, Pridgar, & van der Kooy, 1994). For example, Loy et al. (1993) employed a context discrimination training procedure where rats were poisoned with lithium chloride (LiC1) after they consumed a sucrose solution in a set of boxes (Context A). In contrast, poisoning did not follow sucrose intake in a second set of boxes with different visual, auditory, tactile, and olfactory background cues (Context B). The rats in this experiment gradually came to drink less sucrose in Context A than in Context B.
In the present study on running-based taste aversion, we likewise expected a successful demonstration of context discrimination, not only because many features of Pavlovian conditioning are shared by running- and poison-based taste aversions (see above), but also because some studies suggest a similarity between running and LiCl in the physiological processes causing taste aversion (Dwyer, Boakes, & Hayward, 2008; Nakajima, Urata, & Ogawa, 2006).
The second purpose of the present study was to elucidate the underlying mechanism of context discrimination performance. Researchers disagree on this issue in poison-based taste aversion learning. For instance, in the aforementioned study of Loy et al. (1993), the authors claimed that a direct association between Context A and LiCl must have been added to the taste-LiC1 association to yield the strong avoidance of the target taste in Context A compared with Context B. Their claim is based on the finding that the contexts control led not only the consumption of the target taste but also that of a novel taste solution (Loy et al., 1993, Experiment 1), suggesting that Context A had a direct association with LiCl and thereby took on an aversive character. Related studies demonstrating that a background context can be associated with LiC1 poisoning (e.g., Best, Brown, & Sowell, 1984; Best, Best, & Mickley. 1977; Boakes, Westbrook, & Barnes, 1992; Mitchell & Heyes, 1996; Symonds & Hall, 1997, 1998; Willner, 1978) buttress this interpretation. Another piece of evidence supporting Loy et al.'s claim is that a context extinction treatment reduced the contextual control of the original taste intake (Loy et al., 1993, Experiment 2).
However, other researchers (e.g., Boakes et al., 1997; Nakajima et al., 1995; Skinner et al., 1994) have argued against Loy et al.'s (1993) conclusion. In these studies, the contextual control of LiCl-based taste aversion was unaffected by posttraining extinction treatments. Notably, even in Loy et al.'s research, extinction was not able to completely abolish the contextual control of taste aversion, a finding suggesting another mechanism beyond the direct context-LiC1 association. Furthermore, some experimental studies report that such contextual control does not transfer to novel taste solutions (Archer & Sjoden, 1980; Puente et al., 1988). Given such evidence, those opposing the conclusions of Loy et al. have attributed the critical underlying mechanism of contextual discrimination to the hierarchical modulation (or occasion setting, cf. Schmajuk & Holland, 1998) of the taste-LiC1 association by contexts rather than the summation of associative strengths of taste-LiC1 and context-LiC1 connections. Given the theoretical importance of context discrimination, the present work attempts to assess the mechanism of contextual control in running-based taste-aversion learning.
The major purpose of this study was to demonstrate contextual control of running-based taste aversion in rats. Accordingly, in this experiment, two strains of rats were employed to improve the odds of success. Although we had no initial explicit hypothesiss regarding any strain differences, it was gradually revealed that these two strains differed in contextual discrimination performance. The second purpose of this experiment was to examine the maintenance of context discrimination without further training.
Subjects. Eight experimentally naive male Sprague-Dawley rats (Jbc:SD) and eight experimentally naive male Wistar (Jbc:WI) rats (purchased from a local supplier, Keari Co. Ltd., Osaka, Japan) were housed in individual hanging cages in a vivarium on a 12:12 hr light--dark cycle (lights on at 0800 hr) at approximately 22[degrees]C and 60% humidity. They were maintained on an ad-lib food and water schedule until 9 weeks old, when water was removed from the home cages for the experiment. Thereafter, all rats were allowed to drink tap water for 15 min in the home cages starting at 1310 hr (i.e., 5 hr 35 min before each daily session) throughout the experiment: The limited water was accessible from a needle-pin nozzle protruding through a hole in the back wall of each cage. Food was always available, and the rats' average body weight increased from 332.9 g (range: 306374 g) to 535.4 g (450-620 g) over the 56-day experimental period; there were no strain differences in this measure.
Apparatus. All experimental sessions were administered in an experimental room, where eight activity wheels (15-cm wide, 30 cm in diameter) were arranged in a 4 x 2 pattern on a wall. Each wheel had two side walls made of perforated aluminum sheets and a running surface made of 2-mm metal rods spaced 1 cm apart. A full turn of each wheel was counted automatically by a handcrafted magnetic switch system. The experimental room also had eight stainless steel cages on a table and eight commercial pet den cages on the floor. The stainless steel cages on the table were copies of the hanging home cages (20 cm wide, 25 cm long, 18.7 cm high) made of metal side plates and grid wires. Fluid was provided via a glass bottle with a simple metal spout with a 2-mm hole (KN-671, Natsume Seisakusho Co., Tokyo, Japan) inserted from the wire ceiling. The end of the spout was 16.5 cm above the cage floor. Each of the pet den cages (23 cm wide, 30 cm long, 26 cm high) were made of wire grids coated with white vinyl resin and a green plastic tray (7.6 cm deep) with pine shavings for bedding. A plastic bottle with a ball-bearing spout with a 6-mm hole (WBF-70, Marukan Co. Ltd., Osaka, Japan) was attached to a wire on the side of each cage to provide fluid. The end of the spout was 9.5 cm above the cage floor. When a rat licked the spout, the bouncing ball in the spout made a clanking noise.
Fluorescent ceiling lamps were on in the lighted context (L): the illumination levels around the cages ranged from 165 to 445 lx. In addition, an interband noise at 65 dB (re Scale C) was generated by an FM radio. In the dark context (D), the fluorescent lamps and the radio were turned off, but a table lamp in the corner of the room cast dim illumination of 20 lx around the cages. Rats were carried by 16 individual clear acrylic boxes with opaque flaps (9.5 cm wide, 19.5 cm long, 9.5 cm high) on a cart between the vivarium and the experimental room: the travel time was less than 1 minute.
Procedure. All experimental sessions were conducted with a single squad of 16 rats starting at 1845 hr on successive days. Combinations of the cage types and the background contexts were counterbalanced within each rat strain group. In other words, in the L context, half of the rats drank fluid from the silent bottles in the stainless steel cages, while the remaining rats did so from the noisy bottles in the pet den cages. The allotted cage types were swapped under the D context.
The rats were initially adapted to consuming tap water for 15 min in the cages for 6 days in the order of LDDLLD. On each of the following 20 days, rats were given 15-min access to a salty solution, specifically, a mixture of 7.5 g sodium chloride (NaC1) and 2.3 g monosodium glutamate (MSG) in 1 liter of tap water. After consuming the salty solution, half the rats were allowed to run in the activity wheels for 30 min under the same background context, while the remaining rats were immediately returned to the home cages. This assignment was completely counterbalanced with regard to drinking cage types, background context, and rat strain. Because all rats were treated in a single squad, as noted previously, the order of background contexts over this 20-day discrimination phase was identical for each rat: LLLDDDLLDDLDDLLLDDLD.
The rats were then kept in the vivarium for 14 days with a 30-min access to tap water every day. Then, they were twice given 15 minutes of access (at 1310 hr and 1845 hr) to tap water in the home cages for 2 days to prepare for the 4 days of salty solution retraining in the context order of LDDL.
Analysis. The amount of fluid intake was measured by weighing each bottle before and after drinking using an electric balance to the nearest 0.1 g. Critical comparisons of intakes were assessed by paired t tests with a significant level of p < .05 (two-tailed). When necessary, two-way analyses of variance (ANOVAs), with strain as a between-subject factor and block (or period) as a within-subject factor, were also applied to the data of interest to make statistical decisions with the alpha risk of p < .05.
Results and Discussion
Figure 1 shows the mean ([+ or -] standard error, SE) intake of the salty solution over the 20 contextual discrimination days. Two lines in each panel illustrate the performance of the rats that ran in the L context (solid line) and the performance of the rats that ran in the D context (dotted line). Expression of running-based taste aversion came to be controlled by the background contexts in both rat strains, although the effect was quite small in the Jbc:WI. On the last two training days, consumption was significantly less in the running context than that in the nonrunning context for both the Jbc:SD rats, t(7) = 2.69, p =.031 and the Jbc:WI rats, t(7) = 2.64, p = .034.
The contextual control of running-based taste aversion may be understood more easily by reference to ratios of the form, a/(a + b), where a is the intake of salty solution in the nonrunning context and b is the intake for the running context. A single ratio was calculated from the 2-day (i.e., nonrunning and running days) consumption data for each rat. Because of the characteristics of the training sequence employed in this experiment, the 2 days paired for the ratio calculation are not necessarily successive: Specifically, the paired day numbers were 1 vs. 4, 2 vs. 5, 3 vs. 6, 7 vs. 9, 8 vs. 10, 11 vs. 12, 13 vs. 14, 15 vs. 17, 16 vs. 18, and 19 vs. 20. A ratio of 0.5 indicates no contextual control, while a ratio of 1.0 indicates complete control of running-based taste aversion. As depicted in the circles and triangles of Figure 2, the ratio was numerically larger in the Jbc:SD rats than in the Jbc:WI rats in later blocks. Figure 2 also displays whether the context discrimination is reliable in each block by filling the symbol when the intake is significantly lower in the running context than in the nonrunning context (p < .05, paired t tests).
A 2 (strain) x 10 (block) ANOVA, applied to the ratio data summarized in Figure 2, yielded a significant main effect of block, F(9, 126) = 7.90, p < .001. Although the main effect of strain was not significant, F(1, 14) =1.73, its interaction with block was significant, F(9, 126) = 3.09, p = .002. Subsequent analyses revealed that a significant block effect was observed in the Jbc:SD rats, F(9, 126) = 9.40, p < .001, not in the Jbc:WI rats, F(9, 126) = 1.59. Although the strain difference did not reach significance in any single block, largest F(1, 14) = 4.13, p = .062, inspections of the individual data revealed that one of the Jbc:SD rats showed extraordinarily low ratios over blocks (e.g., it was 0.04 in the last block). Excluding this rat, the Jbc:SD rats would have expressed much better performance as illustrated in the square symbols of Figure 2. A 2 (strain) x 10 (block) ANOVA was subsequently applied to the data of these remaining seven Jbc:SD rats and all eight Jbc:WI rats, yielding significant main effects of group, F(1, 13) = 11.17, p = .005, and block, F(9, 117) = 8.64, p < .001, and their interaction, F(9, 117) = 3.77, p < .001. Subsequent analyses revealed that a significant block effect was observed in the Jbc:SD rats, F(9, 117) = 10.84, p < .001, but not in the Jbc:Wi rats, F(9, 117) = 1.57. The strain difference was significant in the 3rd, 4th, 6th, 7th, 8th, and 10th blocks, Fs(1, 13) > 8.04, ps <.014.
The rats were kept in the home cages without any special treatment for 16 days after the original training. As illustrated in the right portions of Figure 1, this break had little effect on the context discrimination performance of the Jbc:SD rats, but it caused a loss of contextual control for the Jbc:WI rats. Comparing consumption on the first two training days after the break, the intake was marginally lower in the running context than the nonrunning context for the Jbc:SD rats, t(7) = 2.35, p = .051; the contexts had no differential control in the Jbc:WI rats, t < 1.
The same conclusion can be drawn by inspecting Figure 2. Although a 2 (strain) x 2 (phase: pre- vs. postbreak) ANOVA for the data from the two blocks surrounding the break yielded no significant main or interactive effects, a similar ANOVA excluding the aberrant Jbc:SD data yielded significant main effects of strain, F(1, 13) = 9.31, p = .009, and phase, F(1, 13) = 6.57, p = .024. Their interaction, however, was not significant, F(1, 13) = 1.36.
As expected, the mean ([+ or -] SE) number of wheel turns gradually increased over the experimental phase from 36.5 [+ or -] 4.4 to 136.6 [+ or -] 20.4 for the Jbc:SD rats and from 65.6 [+ or -] 8.7 to 134.6 [+ or -] 22.7 for the Jbc:WI rats. A 2 (strain) x 10 (block) ANOVA yielded a significant main effect of block, F(9, 126) = 13.39, p < .001, but the strain had no main, F < 1, or interactive, F (9, 126) = 1.78, effect. Notably, the same patterns were obtained using the data excluding the aberrant Jbc:SD rat data: The main effect of block was significant, F(9, 117) = 11.45, p < .001, while the main effect of strain, F < 1, and the interaction, F(9, 117) = 1.52, were not significant. The amount of running decreased after the 16-day break for the both strains. The mean ([+ or -] SE) numbers of wheel turns of the first and second blocks after the break were 91.4 [+ or -] 16.4 and 76.5 [+ or -] 23.1 for the Jbc:SD rats and 64.8 [+ or -] 18.3 and 88.5 [+ or -] 21.1 for the Jbc:WI rats. A 2 (group) x 2 (phase: pre- vs. postbreak) ANOVA yielded a significant main effect of phase, F(1, 14) = 14.3, p = .002, but the main effect of strain and the interaction were both nonsignificant, Fs < 1. The same conclusion was reached after excluding the problematic Jbc:SD rat: The main effect of phase was significant, F(1, 13) = 16.52, p = .001, while the main effect of strain and the interaction were not significant, Fs < 1.
Experiment 1 successfully demonstrated contextual control of running-based taste aversion, at least in Jbc:SD rats, and this control was well maintained after a break of 16 days. One may, however, see our use of different bottle types (the "silent" glass bottles and the "noisy" plastic bottles) in the corresponding contexts as a problem. Archer and his colleagues (Archer, Sjoden, & Carter, 1979; Archer, Sjoden, Nillson, & Carter, 1980), who also used a similar set of bottles, have reported that stimulation from the bottle spouts was the most salient among the multisensory contextual cues they used. However, Holder (1988a, 1988b) has questioned its role in Archer et al.'s research (see Archer, Sjoden, & Nilsson, 1985; Sjoden & Archer, 1989, for reviews). Essentially, Holder's criticism rests on two points. First, rats can discriminate the tastes arising from glass and plastic bottles (e.g., Garcia, McGowan, & Green, 1972), implying that the contextual control observed in Archer et al.'s research might be ascribed to differential conditioning of two tastes rather than modulatory control of a target taste by two contexts. Second, Holder has asserted that because spout stimulation always accompanied with the target taste, it should not be considered as a background context but rather regarded as a component of the target CS. Holder's criticism of Archer et al.'s research is equally applicable to Experiment 1 of our study, which used a similar set of bottles. Therefore, in Experiment 2, we used a single type of bottle in both contexts to more clearly demonstrate the contextual control of running-based taste aversion.
A second purpose of Experiment 2 was to elucidate the underlying mechanism of contextual control of running-based taste aversion. We attempted to accomplish this purpose with two assessment procedures. First, a two-bottle (target vs. water) choice test was employed in both contexts as a posttraining probe. As noted by Boakes et al. (1997), the direct context--US association would be expected to affect the total amount of fluid intake but not have a conditional effect on the target and water intakes. On the other hand, if the contexts function as modulators (i.e., occasion setters), preference for the water over the target should be greater in the running context than in the nonrunning context. A second procedure to assess the underlying mechanism of contextual control was context exposure after training. As previously noted, studies on LiCl-based taste aversion yielded conflicting results in the effect of context extinction, having little effect in Boakes et al. (1997) and Skinner et al. (1994) but significantly reducing, if not completely abolishing, contextual control in Loy et al. (1993). Notably, we have previously reported in LiCl-based taste aversion that simple exposure to contexts had no effect on the performance, while context exposure with tap water reduced, but did not abolish, the contextual control (Nakajima et al., 1995). By the same token, Experiment 2 investigated the effects of context extinction with and without water presentation on the contextual control of running-based taste aversion.
Subjects. The subjects were experimentally naive male Jbc:SD rats. Their supplier, housing, and preexperimental treatment were the same as in Experiment 1. Jbc:WI rats were not employed here because of the poor performance of this strain shown in Experiment 1.
Although a within-subjects design was employed in this experiment, the number of Jbc:SD rats was doubled here to 16 to retain sensitivity to small effects. All rats were allowed to drink tap water for 15 min in the home cages starting at 1300 hr or 1400 hr (i.e., 5 hr before each daily session) throughout the experiment. The food was always available in the home cages. Their average body weight increased from 353.6 g (range: 319-394 g) to 603.0 g (522-732 g) over the 94-day experimental period.
Apparatus. Like Experiment 1, rats were trained in the experimental room with eight activity wheels. The eight stainless steel cages employed in Experiment 1 were now on the room floors, and they were used for a nonrunning control treatment instead of target taste drinking. Two sets of drinking cages were employed in this experiment. The first set was eight pet den cages used in Experiment 1. They were now on the table, and their trays were covered by pebbles rather than pine shavings. A plastic bottle with a ball-bearing spout (WBF-70, Marukan Co. Ltd., Osaka, Japan) was attached to a wire side of each cage to provide fluid. The end of the spout was 9.5 cm above the cage floor. The second set was eight polycarbonate cages placed on the floor of the room. No bedding materials were in the cages. They had concave wire lids (24 cm wide, 40 cm long, 18 cm high), and fluid was provided by a WBF-70 bottle attached to the lid of each polycarbonate cage with the spout end 14.5 cm above the cage floor. The illumination and sound level conditions were identical to those of Experiment 1, but the combination of the drinking cage types and background contexts was fixed for all rats in this experiment for simplicity. In addition, we attempted to enhance the difference between the contexts by giving transportation cues. Accordingly, rats carried in clear acrylic boxes with opaque flaps (9.5 cm wide, 19.5 cm long, 9.5 cm high) were treated in the pet den cages of the L context. On the other hand, rats carried in tall black acrylic cylinders (9.5 cm wide, 9.5 cm long, 19.5 cm high) were treated in the polycarbonate cages of the D context.
Procedure. All experimental sessions were conducted with two squads of eight rats each: the first squad started at 1800 hr and the second at 1900 hr. The rats were initially adapted to consuming tap water for 15 min in the home cages for two days and then in the experimental room for four days in the order of LDDL or DLLD for the first and second squad of rats, respectively.
Context discrimination. On each of the following 20 days, rats were given 15-min access to the salty solution employed in Experiment 1. After its consumption, rats were either allowed to run in the activity wheels for 30 min or confined in the stainless steel cages for the same duration to equate the amount of context exposure. In this experiment, postconsumption treatment (running or nonrunning) was fixed on a given day, while the physical contexts were changed across the two squads. Specifically, on the running (R) days, the first squad of rats was carried in the clear acrylic boxes to the lighted noisy room (L) where they drank the solution in the pet den cages before running in the activity wheels. The second squad was carried in the black cylinders to the dark silent room (D) where they drank the solution in the polycarbonate cages before running in the activity wheels. On the nonrunning (N) days, the first squad was carried in the black cylinders to the dark silent room (D) where they drank the solution in the polycarbonate cages before confinement in the stainless steel cages. The second squad was carried in the clear acrylic boxes to the lighted noisy room (L) where they drank the solution in the pet den cages before confinement in the stainless steel cages. The treatment order of the 20-day training was NRRRNNNRRNNNRNRRNRRN. In addition to the aforementioned contextual cues, laboratory chow (MF; Oriental Yeast Co., Ltd., Tokyo) was available to the rats in the pet den cages and the polycarbonate cages on the running days in an attempt to facilitate context discrimination.
Two-bottle choice test. On the next day, two bottles spaced 8 cm apart were presented to the rats for 15 min in the running context (L for the first squad, D for the second squad) as a choice preference test between the salty solution and tap water. The left-right positions of the bottles were counterbalanced across rats in each squad. The rats were immediately returned to the vivarium after testing. The same test was administered on the next day in the nonrunning context (D for the first squad, L for the second squad).
Retraining and context extinction without any fluid. After the reinstated discrimination training of four days in the order of RNNR, the laboratory chow was removed from the cages in the experimental room to assess its role in context discrimination. The training was then continued for an additional 14 days in the order of NRRNNRNRNRRNNR. On the following 10 days, the context extinction treatment was administered with the order of NRRNNRNRRN. All rats were kept in the drinking cages for 20 min with the fluid bottles removed. Immediately after returning to the vivarium, they were given 15-min access to tap water in the home cages.
On the next four days, bottles of the salty solution were reintroduced to the drinking cages to assess the effect of context extinction on contextual control of running-based taste aversion. The treatments of the rats were the same as those in the context discrimination training with the order of NRNR.
Context extinction with tap water. All rats were then given a context exposure with tap water for each of the next 30 days with the order of RNNRNRRNNRRRNNRRNNRNNRRNRNRNNR. The treatment was identical to that of the aforementioned context extinction except that a tap water bottle was attached to each cage in the room. The effect of this treatment was assessed by the four-day NRNR training with the salty solution.
Results and Discussion
Context discrimination. As in Experiment 1, the rats acquired running-based taste aversion with the background context gradually coming to control it. The mean ([+ or -] SE) salty solution intakes are shown in Figure 3. Although there were 2 squads of rats with different physical contexts, the data for all rats were collapsed in this experiment because the postconsumption treatment (running or nonrunning) was fixed in a given day. Furthermore, the factor of physical identity was completely counterbalanced across rats, and it had no systematic effect on the fluid consumption. Figure 4 displays the context difference ratios calculated as in Experiment 1. A one-way ANOVA applied to the ratio data of the context discrimination phase (the first section of Figure 4), yielded a significant effect of block, F(9, 135) = 6.29, p < .001, reflecting the acquisition of context discrimination. The filled symbols indicate that intake was significantly lower in the running context than the nonrunning context (p < .05, paired t test). For example, at the end of the 20-day context discrimination training, consumption was statistically lower in the running context than in the nonrunning context, t(15) = 7.79, p < .001, for the last block of two days.
For unknown reasons, the rats of Experiment 2 ran more than those of Experiment 1. The mean ([+ or -] SE) number of wheel turns on the first block was 104.5 [+ or -] 11.7, quickly increased to 210.2 [+ or -] 18.3 by the fourth block, and then hovered around 200 onward. A one-way ANOVA applied to the 10-block data yielded a significant effect of block, F(9, 135) = 13.79, p < .001.
Two-bottle choice test. Figure 5 illustrates the mean ([+ or -] SE) intakes of the salty solution and tap water in the two-bottle testing conducted in the running and nonrunning contexts. It implies the presence of a direct context--US association because the total amount of fluid intake was less in the running context. A 2 (fluid) x 2 (context) ANOVA yielded significant main effects of fluid, F(1, 15) = 37.57, p < .001, and context, F(1, 15) = 14.79, p = .002, but the fluid x context interaction was not significant, F < 1. The null interaction indicates the context had no conditional effect on the salty target and tap water intakes. This was also reflected in the target preference ratio calculated in the form of x/(x + y), where x is the salty solution intake and y is the tap water intake. Although the mean ([+ or -] SE) ratio was slightly less in the running context, 0.14 [+ or -] 0.04, than in the nonrunning context, 0.24 [+ or -] 0.06, this difference was not significant, t(14) = 1.40. These results taken together provide no support for a modulatory property of the background contexts, although this requires some qualification because the context factor was confounded with the order of test days (see the Procedure section).
Retraining. Restatement of the discrimination training resulted in some loss of contextual control of running-based taste aversion, as shown in the beginning of the second section in Figures 3 and 4. Although the difference ratio was numerically well above the chance level of 0.5 in the first block of the retraining phase, the difference in intake between the running and nonrunning contexts was not statistically significant in this block (as indicated by the unfilled symbols). Furthermore, the drop in the ratio from the last (i.e., the 10th) block of the original training was marginally significant, t(15) = 2.12, p = .051. This reduction, however, was only temporary. The rats showed much better performance in the second block of the retraining phase as shown by a significant difference between the running and nonrunning contexts, t(15) = 5.08, p < .001.
From the third block onward, the laboratory chow was removed from the running context to exclude any motivational effect of extra feeding on context discrimination. Inspection of the data of the third block revealed a slight, marginally significant drop from the second block, t(15) = 1.93, p = .073, and the difference between the running and nonrunning contexts was still significant, t(15) = 4.07, p = .001. These results suggest the chow played little if any role in the contextual control observed in the preceding blocks of training. However, discrimination performance deteriorated in the next two blocks followed by recovery to the original level in the succeeding four blocks. Unfortunately, we are unable to offer any clear explanation of this drop in the fourth and fifth blocks.
The mean ([+ or -] SE) number of wheel turns in the first block of the retraining phase was 245.6 [+ or -] 28.7, which was significantly greater than the data of the last (i.e., 10th) block of the original training, 196.6 [+ or -] 23.5, t(15) = 3.28, p = .005. Thereafter, the mean number of wheel turns returned to the original level of around 200. A one-way ANOVA applied to the nine-block data yielded a significant effect of block, F(8, 120) = 6.17, p < .001.
Effect of context extinction without any fluid. Exposures to the contexts without any fluid for 10 days (five days each for the running and nonrunning contexts) had little effect as shown in the first block of the postextinction phase (the third sections of Figures 3 and 4). The difference between the running and nonrunning contexts was significant in this block, t(15) = 5.50, p < .001. Furthermore, difference ratios were statistically equivalent between the last (i.e. ninth) block of the preextinction phase and the first block of the postextinction phase, t < 1, although we observed a small, nonsignificant, drop of performance in the second block, t(15) = 1.72. These results imply a hierarchical modulation of taste aversion by the contexts; this is inconsistent with the conclusion drawn from the two-bottle test.
The mean ([+ or -] SE) number of wheel turns significantly dropped, t(15) = 4.83, p < .001, from the last (i.e., ninth) block of the preextinction phase, 195.6 [+ or -] 21.2, to the first block of the postextinction phase, 146.6 [+ or -] 16.3. The same level of running was observed in the second block, 151.3 [+ or -] 14.6, t < 1.
Context extinction with tap water and its effect on taste aversion. The fourth section of Figure 3 depicts the intake data when the rats were exposed to the two contexts with tap water in the absence of running. This procedure provided tap water transfer data for contextual control. The contexts had a differential effect on water intake (see Figure 4). The water intake was significantly lower in the running context compared with the nonrunning context in both the first two-day block, t(15) = 3.06, p = .008, and the following 6 blocks, ts(15) > 2.64, Ps < .019. The difference was also significant in the 9th (t = 2.64, p = .019), 12th (t = 3.53, p = .003), and 15th (or final) blocks (t = 2.14, p = .048). This long-lasting control of water intake by context was unexpected, although the difference ratio gradually decreased over the blocks (see Figure 4), as supported by a one-way ANOVA, F(14, 210) = 3.38, p < .001. Notably, this experiment had a large enough number of subjects (n = 16) to uncover otherwise small differences between the running and nonrunning contexts.
The 30-day (15-block) exposures to the contexts with tap water had little effect on the contextual control of the salty solution. Although the intake difference between the running and nonrunning contexts failed to reach significance in the first block of testing (t < 1), comparison of the difference ratio between the last blocks of discrimination training and the first block of testing was also nonsignificant, t < 1. Furthermore, the intake difference was significant in the second Nock of testing, r(15) = 2.88, p = .011. In addition, the mean ratio averaged over the two blocks of testing also did not differ from the mean ratio averaged over the last two blocks of the preceding discrimination training. These results indicate a modulatory 1 unction of context. One may argue against this statement by noting that the direct context--US association was not completely extinguished after the 30-day exposures to the contexts, as suggested by the significant context difference at the end of the 30-day treatment of context exposures. Yet, it is difficult to see how this argument can be justified in light of the full maintenance of contextual control of the salty solution after the significant reduction of contextual control of tap water intake over the 15 blocks of context extinction.
The primary purpose of the present study was to demonstrate the contextual control of running-based taste aversion in rats, accomplished in both Experiments 1 and 2. Experiment 1 showed that Jbc:SD rats quickly learned to differentiate consumption of a salty solution between the running and nonrunning contexts, while Jbc:WI rat performance barely reached a significance level of p < .05 at the end of 20-day training. These two strains were equivalent in their age, body weight, and amount of running. We have no clear explanation of the strain difference in the observed discrimination performance. It is unlikely that Jbc:WI rats are unsuitable subjects for running-based taste aversion learning, because we have repeatedly observed this kind of learning with Jbc:WI rats (e.g., Hayashi, Nakajima, Urushihara, & Imada, 2002; Masaki & Nakajima, 2006; Nakajima, 2004, 2008; Nakajima et al., 2000, 2006). In addition, our unpublished research (Nakajima, 2011) has shown that Jbc:SD and Jbc:WI rats are equally able to acquire a running-based taste aversion to the taste solution employed here (a NaC1 + MSG mixture). Furthermore, Jbc:WI rats in our previous research have shown modest performance on the contextual control of LiCl-based taste aversion (Nakajima et al., 1995), although the target solution and the background cues employed in that study differed from those in the present study. Future research should clarify why the combination of contextual learning and running-based taste aversion has a disadvantage for Jbc:WI rats.
The good performance of the Jbc:SD rats was well maintained over a 16-day break after the original training. This is consistent with Boakes et al. (1997), who reported that contextual control of LiCl-based taste aversion was maintained 15 to 16 days after the training. Thus, our finding was not without precedent and provides an additional correspondence between LiCl-based and running-based taste aversions.
Contextual control of running-based taste aversion was also successfully demonstrated in the Jbc:SD rats in Experiment 2, where differential bottle cues were not employed, thus avoiding the criticism offered by Holder (1988a, 1988b) that bottle stimulation should not be treated as a contextual cue. Experiment 2 also provides some hints about the underlying mechanism of contextual control of running-based taste aversion. The two-bottle choice test administered after the original training suggests that the direct context-running is a major factor in the contextual control observed here, because the total fluid intake was less in the running than in the nonrunning contexts, while the target preference ratio was equivalent between the contexts. The direct context-running association is also reflected in the transfer to tap water intake. These results taken together indicate that the running context turns out to be aversive by the direct context-running association. Notably, two related studies buttress this claim. First, Masaki and Nakajima (2008) have demonstrated that confinement in a distinctive chamber followed by wheel running endows rats with a conditioned avoidance of that chamber. Second, Masaki (2011) has reported that rats having running-based saccharin aversion in training boxes showed transfer of attenuated drinking to tap water in the same boxes.
Nevertheless, this indication is not consonant with the maintenance of contextual control of running-based taste aversion after exposures to the contexts without any fluid exposures aimed to extinguish the direct context-running association. The contextual control was also well maintained after context exposures with tap water. These results signify the modulatory function of the contexts, because a modulator is highly resistant to extinction by simple exposure (e.g., Bouton & Swartzentruber, 1986; Holland, 1989a, 1989b; Rescorla, 1985, 1986). According to the modulatory account, transfer of contextual control is possible if the original and transfer stimuli share the perceptually common element or if the transfer stimulus has been trained in a similar way. The latter is unlikely because the transfer stimulus is tap water in the present study. The former possibility deserves consideration, however, because tap water was also presented to the rats using the same types of bottles and spouts as that for the original taste solution. In other words, the bottle/spout cue must have formed a portion of the target stimulus (Holder, 1988a, 1988b). Hence, the contextual control would transfer, by stimulus generalization, to intake of the tap water provided by the same bottles/spouts. The validity of this account remains to be explored in future studies. If this is the case, perhaps the most plausible, if somewhat cumbersome, explanation would be that contexts play two roles in running-based taste aversion: direct association with a running US and modulatory control of the taste-running association. We, however, assert that modulation was a secondary, rather than a primary, factor of the contextual control demonstrated in the present study, because the contexts had no conditional effect on the salty target and tap water intakes in two-bottle choice testing but did have a simple effect on total consumption.
Furthermore, the maintenance of contextual control after context exposure does not necessarily entail the modulatory function of contexts, because there are some other processes that make it possible. For example, as briefly noted by Loy et al. (1993), contexts may acquire not only excitatory but also inhibitory properties. Specifically, when a CS is paired with a US in Context A but not in Context B, rats may establish an inhibitory association between Context B and the US as well as an excitatory Context A--US association. A number of studies show that simple exposures to a conditioned inhibitor are not enough to diminish the inhibitory property, and, if anything, it increases inhibition (e.g., DeVito & Fowler, 1986, 1987; Hallam, Grahame, Harris, & Miller, 1992; Williams, Travis, & Overmier, 1986; Witcher & Ayers, 1984; Zimmer-Hart & Rescorla, 1974). Accordingly, the well-maintained contextual control over taste aversion after context exposure may reflect a sustained inhibitory property of the nonrunning context.
Another possible account of the maintained contextual control is provided by configural conditioning theories (e.g., Kehoe, 1988; Kehoe & Gormezano, 1980; Pearce, 1987). According to these theories, a unique configuration of a stimulus compound, rather than its elements, is associated with a US. In the present study, where a target taste solution (T) in Context A was followed by running while T in Context B was not, the task given to rats is regarded as the differential conditioning of Stimulus TA and Stimulus TB. Subsequent exposures to the contexts should have no effect on this differential conditioning if Stimuli TA and TB are completely unique from Stimuli A and B, respectively. The amount of this effect depends on the degree of generalization between these stimuli. For example, if we calculate the associative strengths of Stimuli TA and TB according to the mathematical model of configuration/generalization proposed by Pearce (1987), their changes would be small. Assuming for simplicity that all stimulus elements had the same level of salience and that the original learning and subsequent extinction treatment had been complete at the final session of each phase, the associative strength of Stimulus TA would change from 1 to 0.73 and that of Stimulus TB from 0 to 0.07. In other words, we would still have a substantial difference of 0.66 in the associative strength after the context extinction treatment. Parenthetically, the difference would be 0.73 if we adopt a modified Pearce model (Wilson & Pearce, 1989) incorporating the rule to handle the resistance of an inhibitor to extinction, because it predicts an unchanged associative strength for Stimulus TB.
These two processes--inhibitory conditioning of the nonrunning context and configural conditioning--can account for not only the maintenance of contextual control after context exposure but also for the results of the two-bottle choice testing and the transfer to tap water. Therefore, they are more parsimonious explanations than the aforementioned two-role position, attributing the contextual control observed in the present study to the direct excitatory conditioning of the running context and the complementary modulatory properties of the contexts. Unfortunately, the present study was not designed to test these explanations.
Although the underlying psychological mechanism of contextual control has not been definitively established in the present study, the demonstration of contextual control of running-based taste aversion does informatively expand the field of research. For example, in a series of investigations, Skinner and colleagues showed that LiCl-based taste aversion can be controlled by not only external background contexts but also internal contexts induced by drugs such as morphine (see Skinner, Goddard, & Holland, 1998, for a review). If the underlying physiological process for establishing taste aversion learning is shared by LiC1 and running (Dwyer et. al., 2008; Nakajima et al., 2006), drug states would also function as contexts for running-based taste aversion.
This study was financially supported by JSPS KAKENHI Grants (C-21530779, C-24530931) to the second author, who supervised the experiments administered by the first author in partial fulfillment of requirements for the MA degree. This research project and the animal facility were approved by the Animal Care and Use Committee of KGU in 2009, based on the Guide for the Care and Use of Laboratory Animals published by the National Academy of Sciences of the Unites States in 1996, as well as on Japanese law (the Act on Welfare and Management of Animals) and the guideline published by the Science Council of Japan (the Guidelines for Proper Conduct of Animals Experiments) in 2006.
Correspondence concerning this article should be addressed to Sadahiko Nakajima, Department of Psychological Science, Kwansei Gakuin University, Nishinomiya 662-8501, Japan. E-mail: email@example.com
ARCHER, T., & SJODEN, P.-0. (1980). Context-dependent taste-aversion learning with a familiar conditioning context. Physiological Psychology, 8, 40-46.
ARCHER, T., SJODEN, & CARTER, N. (1979). Control of taste aversion extinction by exteroceptive cues. Behavioral and Neural Biology, 25, 217-226. doi:10.1016/S0163-1047(79)90571-5
ARCHER, T., SJODEN, & NILSSON, L.-G. (1985). Contextual control of taste-aversion conditioning and extinction. In P. D. Balsam & A. Tomie (Eds.), Context and learning (pp. 225-271). Hillsdale, NJ: Erlbaum.
ARCHER, T., SJODEN, P.-0., NILSSON, L.-G., & CARTER, N. (1980). Exteroceptive context in taste-aversion conditioning and extinction: Odour, cage, and bottle stimuli. Quarterly Journal of Experimental Psychology, 32, 197-214. doi:10.1080/14640748008401157
BEST, M. R., BROWN, E. R., & SOWELL, M. K. (1984). Taste-mediated potentiation of noningestional stimuli in rats. Learning and Motivation, 15, 244-258. doi:10.1016/0023-9690(84)90021-3
BEST, P. J., BEST, M. R., & MICKLEY, G. A. (1977). Conditioned aversion to distinct environmental stimuli resulting from gastrointestinal distress. Journal of Comparative and Physiological Psychology, 85, 250-257. doi:10.1037/h0035010
BOAKES, R. A., & NAKAJIMA, S. (2009). Conditioned taste aversions based on running or swimming. In S. Reilly & T. R. Schachtman (Eds.), Conditioned taste aversion: Behavioral and neural processes (pp. 159-178). New York, NY: Oxford University Press.
BOAKES, R. A., WESTBROOK, R. F., & BARNES, B. W. (1992). Potentiation by a taste of a toxicosis-based context aversion: Effect of test solution. Quarterly Journal of Experimental Psychology, 45B, 303-325. doi:10.1080/14640749208401008
BOAKES, R. A., WESTBROOK, R. F., ELLIOTT, M., & SWINBOURNE, A. L. (1997). Context dependency of conditioned aversions to water and sweet tastes. Journal of Experimental Psychology: Animal Behavior Processes, 23, 56-67. doi:10.1037/0097-7403.23.1.56
BONARDI, C., HONEY, R. C., & HALL, G. (1990). Context-specificity of conditioning in flavor-aversion learning: Extinction and blocking tests. Animal Learning & Behavior, 18, 229-237. doi:10.3758/BF03205280
BOUTON, M. E., & SWARTZENTRUBER, D. (1986). Analysis of the associative and occasion-setting properties of contexts participating in a Pavlovian discrimination. Journal of Experimental Psychology: Animal Behavior Processes, 12, 333-350. doi:10.1037/0097-7403.12.4.333
DEVITO, P. L. & FOWLER, H. (1986). Effects of contingency violations on the extinction of a conditioned fear inhibitor and a conditioned fear excitor. Journal of Experimental Psychology: Animal Behavior Processes, 12, 99-115. doi:10.1037/0097-7403.12.2.99
DEVITO, P. L., & FOWLER, H. (1987). Enhancement of conditioned inhibition via an extinction treatment. Animal Learning & Behavior, 15, 448-454. doi:10.3758 /BF03205055
DWYER, D., BOAKES, R. A., & HAYWARD, S. (2008). Reduced palatability in lithium- and activity-based, but not in amphetamine-based, taste aversion learning. Behavioral Neuroscience, 122, 1051-1060. doi:10.1037/a0012703
GARCIA, J., MCGOWAN, B. K., & GREEN, K. F. (1972). Biological constraints on conditioning. In A. H. Black & W. F. Prokasy (Eds.), Classical conditioning II: Current research and theory (pp. 21-43). New York, NY: Appleton.
HALLAM, S. C., GRAHAME, N. J., HARRIS, K., & MILLER, R. R. (1992). Associative structures underlying enhanced negative summation following operational extinction of a Pavlovian inhibitor. Learning and Motivation, 23, 43-62. doi:10.1016/0023-9690 (92)90022-E
HAYASHI, H., NAKAJIMA, S., URUSHUHARA, K., & IMADA, H. (2002). Taste avoidance caused by spontaneous wheel running: Effects of duration and delay of wheel confinement. Learning and Motivation, 33, 390-409.
HETH, C. D., INGLIS, P., RUSSELL, J. C., & PIERCE, W. D. (2001). Conditioned taste aversion induced by wheel running is not due to novelty of the wheel. Physiology & Behavior, 74, 53-56. doi:10.1016/S0031-9384(01)00553-4
HOLDER, M. a (1988a). Possible role of confounded taste stimuli in conditioned taste aversions. Animal Learning & Behavior, 16, 231-234. doi:10.3758/BF03209070
HOLDER, M. D. (1988b). Reply to Sjoden and Archer. Animal Learning & Behavior, 16, 240-241. doi:10.3758/BF03209072
HOLLAND, P. C. (1989a). Feature extinction enhances transfer of occasion setting. Animal Learning & Behavior, 17, 269-279. doi:10.3758/BF03209799
HOLLAND, P. C. (1989b). Occasion setting with simultaneous compounds in rats. Journal of Experimental Psychology: Animal Behavior Processes, 15, 183-193. doi:10.1037/0097-7403.15.3.183
HUGHES, S., & BOAKES, R. A. (2008). Flavor preferences produced by backward pairing with wheel running. Journal of Experimental Psychology: Animal Behavior Processes, 34, 283-293. doi:10.1037/0097-7403.34.2.283
KEHOE, E. J. (1988). A layered network model of associative learning: Learning-to-learn and configuration. Psychological Review, 95, 411-433. doi:10.1037/0033-295X .95.4.411
KEHOE , E. J., & GORMEZANO , I. (1980). Configuration and combination laws in conditioning with compound stimuli. Psychological Bulletin, 8Z 353-378. doi:10.1037/0033-2909.87.2.351
LETT, B. T., & GRANT, V. L. (1996). Wheel running induces conditioned taste aversion in rats trained while hungry and thirsty. Physiology & Behavior, 59, 699-702. doi :10.1016/0031-9384(95)02139-6
LOY, I., ALVAREZ, R., REY, V., & LOPEZ, M. (1993). Context-US associations rather than occasion setting in taste aversion learning. Learning and Motivation, 24, 55-72. doi:10.1006/1mot.1993.1004
MASAK I. T. (2011, September 8). Formation of a context-running association in running-based conditioned taste aversion in rats [in Japanese]. Paper presented at the 71st Annual Meeting of the Japanese Society for Animal Psychology, Tokyo.
MASAK I. T., & NAKAJIMA. S. (2006). Taste aversion induced by forced swimming, voluntary running, forced running, and lithium chloride injection treatments.
Physiology & 88, 411-416. doi:10.1016/j.physbeh.2006.04.013 MASAKI, T., & NAKAJIMA, S. (2008). Forward conditioning with wheel running causes place aversion in rats. Behavioural Processes, 79, 43-47. doi:10.1016/j.beproc .2008.04.006
MITCHELL, C., & HEYES, C. (1996). Simultaneous overshadowing and potentiation of taste and contextual cues by a second taste in toxicosis conditioning. Learning and Motivation, 27, 58-72. doi:10.1006/1mot.1996.0004
NAGAISHI, T., & NAKAJIMA, s. (2010). Overshadowing of running-based taste aversion learning by another taste cue. Behavioural Processes, 83, 134-136. doi:10.1016/j.beproc.2009.11.003
NAKAJIMA, s. (2004). Conditioned ethanol aversion in rats induced by voluntary wheel running, forced swimming, and electric shock: An implication for aversion therapy of alcoholism. Integrative Physiological and Behavioral Science, 32, 95-104. doi:10.1007/BF02734275
NAKAJIMA, S. (2008). Effect of extra running on running-based taste aversion in rats. Behavioural Processes, 78, 470-472. doi:10.1016/j.beproc.2008.01.018
NAKAJIMA, S. (2011, November 4). Running-based taste aversion learning in Wistar, Sprague-Dawley, Fischer, and Lewis rats. Paper presented at the 52nd Annual Meeting of the Psychonomic Society, Seattle, WA.
NAKAJIMA, S., HAYASHI, H., & KATO, T. (2000). Taste aversion induced by confinement in a running wheel. Behavioural Processes, 49, 35-42. doi:10.1016/S0376-6357(00)00071-1
NAKAJIMA, S., KOBAYASHI, Y., & IMADA, H. (1995). Contextual control of taste aversion in rats: The effects of context extinction. The Psychological Record, 45, 309-318.
NAKAJIMA, S., URATA, T., & OGAWA, Y. (2006). Familiarization and cross-familiarization of wheel running and LiC1 in conditioned taste aversion. Physiology & Behavior, 88, 1-11. doi:10.1016/j.physbeh.2006.02.006
PEARCE , J. M. (1987). A model of stimulus generalization for Pavlovian conditioning. Psychological Review, 94, 61-73. doi:10.1037//0033-295X.94.1.61
PIERCE, W. D., & HETH, C. D. (2010). Blocking of conditioned taste avoidance induced by wheel running. Behavioural Processes, 83, 41-47. doi:10.1016/j.beproc.2009.09.005
PUENTE, G. P., CANNON, D. S., BEST, M. R., & CARRELL, L. E. (1988). Occasion setting of fluid ingestion by contextual cues. Learning and Motivation, 19, 239-253. doi:10.1016/0023-9690(88)90003-3
RESCORLA, R. A. (1985). Conditioned inhibition and facilitation. In R. R. Miller & N. E. Spear (Eds.), Information processing in animals: Conditioned inhibition (pp. 299-326). Hillsdale, NJ: Erlbaum.
RESCORLA , R. A. (1986). Extinction of facilitation. Journal of Experimental Psychology: Animal Behavior Processes, 12, 16-24. doi:10.1037/0097-7403.12.1.16
SALVY, S-J., PIERCE, W. D., HETH, D. C., RUSSELL, J. C. (2002). Pre-exposure to wheel running disrupts taste aversion conditioning. Physiology & Behavior, 76, 51-56. doi:10.1016/S0031-9384(02)00687-X
SCHMAJUK, N. A., & HOLLAND, P. C. (Eds.). (1998). Occasion setting: Associative learning and cognition in animals. Washington, DC: American Psychological Association. doi:10.1037/10298-000
SJODEN, P.-0., & ARCHER, T. (1989). Taste-aversion conditioning: The role of contextual stimuli. In T. Archer & L.-G. Nilsson, (Eds.), Aversion, avoidance, and anxiety: Perspectives on aversively motivated behavior (pp. 87-120). Hillsdale, NJ: Erlbaum.
SKINNER, D. M., GODDARD, M. J., & HOLLAND, P. C. (1998). What can nontraditional features tell us about conditioning and occasion setting? In N. A. Schmajuk & P. C. Holland (Eds.), Occasion setting: Associative learning and cognition in animals (pp. 113-144). Washington, DC: American Psychological Association. doi:10.1037/10298-004
SKINNER, D. M., MARTIN, G. M., PRIDGAR, A., & VAN DER KOOY, D. (1994). Conditional control of fluid consumption in an occasion setting paradigm is independent of Pavlovian associations. Learning and Motivation, 25, 368-400. doi:10.1006/lmot.1994.1019
SPARKES, S., GRANT, V. L., & LETT, B. T. (2003). Role of conditioned taste aversion in the development of activity anorexia. Appetite, 41, 161-165. doi:10.1016/S0195-6663(03)00057-6
SYMONDS, M., & HALL, G. (1997). Contextual conditioning with lithium-induced nausea as the US: Evidence from a blocking procedure. Learning and Motivation, 28, 200-215. doi:/10.1006/1mot.1996.0958
SYMONDS, M., & HALL, G. (1998). Is fluid consumption necessary for the formation of context-illness associations? An evaluation using consumption and blocking tests. Learning and Motivation, 29, 168-183. doi:10.1006/1mot.1997.0998
WILLIAMS, D. A., TRAVIS, G. M., & OVERMIER, J. B. (1986). Within-compound associations modulate the relative effectiveness of differential and Pavlovian conditioned inhibition procedures. Journal of Experimental Psychology: Animal. Behavior Processes, 12, 351-362. doi:10.1037/0097-7403.12.4.351
WILLNER, J. A. (1978). Blocking of a taste aversion by prior pairings of exteroceptive stimuli with illness. Learning and Motivation, 9, 125-140.
WILSON, P. N., & PEARCE, J. M. (1989). A role for stimulus generalization in conditional discrimination learning. Quarterly Journal of Experimental Psychology, 41B, 243-273. doi:10.1080/14640748908401195
WITCHER, E. S., & AYRES, J. J. B. (1984). A test of two methods for extinguishing Pavlovian conditioned inhibition. Animal Learning & Behavior, 12, 149-156. doi:10.3758/BF03213134
ZIMMER-HART, C. L., & RESCORLA, R. A. (1974). Extinction of Pavlovian conditioned inhibition. Journal of Comparative and Physiological Psychology, 86, 837-845. doi:10.1037/h0036412
Aya Hashimoto and Sadahiko Nakajima
Kwansei Gakuin University
|Printer friendly Cite/link Email Feedback|
|Author:||Hashimoto, Aya; Nakajima, Sadahiko|
|Publication:||The Psychological Record|
|Date:||Sep 22, 2013|
|Previous Article:||Membership of defined responses in stimulus classes.|
|Next Article:||Risk taking for sexual versus monetary outcomes using the Balloon Analogue Risk Task.|