Printer Friendly

A factor affecting stimulus classes formation in concurrent discriminations in rats.

There are two main views on the formation mechanism of stimulus classes: The one is categorization processes based on simple similarity between stimuli (Bhatt & Wasserman, 1989; Fersen & Lea, 1990; Vaughan & Herrstein, 1987), the other is stimulus associations on the basis of reinforcement concordance (Delius, Ameling, Lea, & Staddon, 1995; Edwards, Jagielo, Zentall, & Hogan, 1982; Nakagawa, 1986, 1992, 1998; Urcuioli, Zentall, Jackson, Smith, & Steirn, 1989, 1998). However, it is important to note that such stimulus associations mechanism as postulated in Nakagawa (1986, 1992, 1998) and Delius et al. (1995) is not necessarily in conflict with categorization processes based on simple similarity between stimuli. Conversely, it seems likely that in nature both principles may often act in consonance. A specific question, however, remains. What precise conditions favor the formation of stimulus-stimulus associations (i.e., stimulus classes), for example? Whether or not the rats can form interstimulus associations on a basis of the same response following the same consequence? This is a very important and ever fundamental issue in behavior analysis. This problem has received far too little experimental attention.

Edwards et al. (1982) have reported that pigeons can acquire an arbitrary stimulus class. In this experiment, pigeons were given two stimulus matching-to-sample tasks: shape matching-to-sample task on some trials and hue matching-to-sample task on other trials in a random order within session, in which a correct match of one hue and one shape was followed by corn and correct match of the other hue and the other shape was followed by wheat. On the transfer test, when hue samples were presented with shape comparisons and vice versa, pigeons chose the comparison associated with the same outcome as the sample. This finding indicates that a stimulus class consists of arbitrary stimuli, related to each other, through their association with a same outcome in pigeons.

Urcuioli et al. (Experiment 2, 1989), using a design analogous to that used by Spradlin, Cotter, and Baxley (1973), trained pigeons on a conditional discrimination involving hue and line samples in Phase 1 (e.g., red-vertical line and green-horizontal line associations). In Phase 2, pigeons were trained on a second task involving the hue samples from Phase 1 and a new pair of comparison stimuli (e.g., red-vertical line and green-horizontal line associations). In Phase 3 test, pigeons matched the line samples from Phase 1 to the comparisons from Phase 2. Urcuioli et al. (1989) found that when the Phase 3 associations were consistent with the hypothesized stimulus classes (e.g., red-vertical line and green-horizontal line associations), pigeons performed well above chance, whereas when the associations were inconsistent with the stimulus classes (e.g., red-horizontal line and green-vertical line associations), pigeons' performance levels were below chance. This finding indicates that pigeons form stimulus classes on a basis of reinforcement concordance.

Using a whole-partial reversal procedure which compared rats' performances on whole reversal sessions (both stimulus pairs reversed, from A+C-, B+D- to A-C+, B-D+ for example) with those on partial reversal sessions (only one pair reversed, from A-C+, B-D+ to A+C-, B-D+ for example), Nakagawa (1978, 1986, 1992, 1998) has shown that rats form stimulus classes between the discriminative stimuli on the basis of the same response during overtraining in concurrent discriminations. In Nakagawa's experiments, rats were trained to criterion or overtrained on two concurrent discriminations (A+C-, B+D-for example) in Phase 1 training in both a simultaneous (Nakagawa, 1986, 1998) and a go-no go successive concurrent discriminations (Nakagawa, 1992, 1998). After completing Phase 1 training, rats received either partial reversal (only one pair reversed from A+C-, B+D-to either A-C+, B+D- or A+C-, B-D+ for example) or whole reversal (both stimulus pairs reversed, from A+C-, B+D- to A-C+, B-D+ for example). The rats for which both discriminations were reversed took fewer days to learn their reversed task than those for which only one discrimination of the two tasks was reversed after overtraining, but not after if reversal occurred immediately upon reaching the criterion in the original learning. Overtraining facilitated the whole reversal, whereas it retarded the partial reversal. Apparently, these findings make it clear that stimulus associations between discriminative stimuli with the same outcome can be formed following overtraining in rats.

Zentall, Steirn, Sherburne, and Urcuioli (1991), using the whole-partial reversal procedure reported by Nakagawa (1986), showed that after many-to-one overtraining in which red (R) and vertical-line (V) samples were associated with a circle (C) comparison and green (G) and horizontal-line (H) samples were associated with a dot (D) comparison, pigeons acquired their reversal learning faster when both sets of associations were reversed than when only one set was reversed. Zentall, Sherburne, Steirn, Randall, Roper, and Urcuioli (1992), using the whole-partial reversal procedure, showed that after overtraining one-to-many conditional discriminations in which a circle (C) sample was associated with green (G) and vertical-line (V) comparisons and a dot (D) sample was associated with red (R) and horizontal-line (H) comparisons, pigeons acquired their reversal learning faster when both sets of associations were reversed than when only one set was reversed. Findings of these two studies make it clear that pigeons form stimulus classes.

Delius et al. (1995), using the whole-partial reversal procedure in their Experiment 3, have shown that pigeons take fewer trials to reach near-asymptote performance (85.5%, correct) in the full-reversal condition than in the half-reversal condition, and that additional prereversal training promotes the occurrence of the full-reversal versus half-reversal advantage in pigeons.

Dube, Callahan, and McIlvane (1993), using auditory stimuli and the serial concurrent reversal technique reported by Vaughan (1988), showed that some rats were capable of serial reversals of concurrent discriminations. This finding suggests that two or more stimuli that are concurrently discriminative for a given consequence (reinforcement or extinction) may constitute a class of functionally equivalent stimuli in rats.

Hall, Ray, and Bonardi (1993) have demonstrated that stimuli shared a common antecedent would come to be treated as equivalent using conditioned suppression procedure. In Experiment 1, rats experienced two stimuli (A and N) each preceded by the same event (food) or by different events (food preceded one but not the other). Stimulus N was then paired with shock, and the generalization of conditioned suppression to A was assessed. Generalization was more marked when A and N had been experienced along with a common antecedent. In Experiment 2, three stimuli (A, B, and N) were presented in initial training. For the first group, A and N were preceded by food and B was not; for a second group A alone was preceded by food. In each group, suppression generalized more readily from N to the stimulus that had received the same initial training as had been given to N.

Nakagawa (1986, 1992), Zentall, Steirn, et al. (1991), Zentall, Sherburne, et al., (1992), Delius et al. (1995), Dube et al. (1993), and Hall et al. (1993) suggest that rats and pigeons can form stimulus classes on a basis of reinforcement concordance.

Nakagawa (1978, 1986, 1992) has asserted that, during the original discrimination training, rats learn a connection between a positive stimulus and an approach response as well as a connection between a negative stimulus and an avoidance response for each discrimination task. But during overtraining they also form associations between the discriminative stimuli on a basis of the same response following the same consequence in concurrent discriminations.

In spite of the reliability that pigeons and rats form stimulus classes, it remains unclear whether or not rats form stimulus classes on a basis of not only reinforcement concordance, say of reinforcement or nonreinforcement, but also the same response (i.e., shared common response) following the same consequence during overtraining in conditional concurrent discriminations. However, no studies have reported that rats form stimulus classes on a basis of the same response during overtraining. The present experiments were conducted to determine whether or not rats form stimulus classes on a basis of the same response in conditional concurrent discriminations.

Experiment 1

Experiment 1 was conducted to investigate whether or not rats formed stimulus classes on a basis of the same response (i.e., shared common response) following the same consequence. Rats were concurrently trained to discriminate four simple stimuli (A, B, C, D for example) where responses to two stimuli were rewarded (A+B+ for example) and responses to the other stimuli were not rewarded (C-D- for example) to reach a criterion or overtrained in a straight runway alley. After completing the original training, they were transferred to conditional successive discrimination learning in a Y maze, in which half of the rats were required to choose the right goal box if the original positive stimulus (either A+ or B+) was presented on the entrance door of each goal box, and they were required to choose the left goal box if the original negative stimulus (either C- or D-) was presented (consistent condition; Group C). The remaining rats were required to choose the right goal box if either one of the original positive stimulus (A+) or one of the original negative stimulus (C-) was presented on the entrance door of each goal box, and they were required to choose the left goal box if either the other of the original positive stimulus (B+) or the other of the original negative stimulus (D-) was presented (inconsistent condition; Group IC). If rats formed stimulus classes on a basis of the same response during overtraining, as the cue-associations theory (Nakagawa, 1986, 1992) postulated, rats trained in the consistent condition should then learn transfer more rapidly than those trained in the inconsistent condition after overtraining, but not after criterion training

Method

Subjects. Twenty-four experimentally naive male Sprague-Dawley rats were used. They were about 210 days old with an initial average body weight of 543 g. The animals were handled for 5 min a day for 12 days. They were maintained on a daily 2-hr feeding schedule prior to the experiment. The amount of food in the daily ration was gradually reduced until the body weight of each animal reached 80% of the baseline weight at the start of the experiment. Water was always available for the animals in their individual home cages. The animals were maintained on a 6:18-hr light:dark cycle with light off at 10:00 p.m.

Apparatus. Both a modified straight runway shown in Figure 1.1 and the same modified Y maze shown in Figure 1.2 were used. The straight runway consisted of a runway with a startbox and a goal box. A guillotine door was located at the front of the startbox. At the entrance of the goal box a piece of cardboard was placed which served as a discriminandum. A gap over which the animals had to jump was located 20 cm in front of the goal box. The apparatus was painted medium-gray inside and lit throughout the experiment by a 10-W fluorescent lamp suspended 45 cm above the top of the runway. Separate starting time and running time were obtained on each trial by means of two electrical digital timers. When the experimenter raised the guillotine door, a microswitch activated the first timer. When the animals interrupted a photobeam gate at 7 cm down the runway, this timer stopped, and at the same time the second timer started. When the animals interrupted the second photobeam gate at 67 cm farther down the runway, the second timer stopped. A running time is the time between the interrupting of the first and second photobeam gates. That is, the distance for running time was 60 cm.

The startbox in this Y maze measured 15 cm in height, 12 cm in width, and 25 cm in length. The distance from the startbox to the bifurcation was 65 cm. The arms of the Y maze were 85 cm in length. At the end of each arm was a goal box. A guillotine door was located at the front of the start box. At the entrance of each goal box a piece of cardboard was placed which served as a discriminandum. A gap over which the animals had to jump was located 20 cm in front of the goal box. The apparatus was lit throughout the experiment by two 10-W fluorescent lamps suspended 45 cm above the top of both the arms. This apparatus was painted medium-gray inside.

Stimuli. Stimulus cards were 12-cm squares of cardboard. Each square was presented at the entrance of each goal box and served as an entrance door. They were arranged so that on rewarded trials the card serving as the correct door could be pushed down easily, thus permitting the animals to gain entrance into the goal box, whereas on nonrewarded trials the card denoting the incorrect door was locked. Vertically striped and horizontally striped cards were used for a vertical-horizontal stripe discrimination. Striped cards had alternating black and white lines 1 cm in width. For a triangle-circle discrimination an equilateral triangle with 10-cm sides and a circle with a diameter of 7.5 cm were used.

Procedure. The animals were given pretraining for 8 days prior to the beginning of discrimination training. On Day 1 the animals were allowed to explore the apparatus for two periods of 7 and 5 min. From Day 2 to Day 4 they were trained to push down a stimulus card and to enter the goal box to obtain food for 10 daily trials. The gap was not present for this stage of the experiment. From Day 5 to Day 8 they were trained to jump over the gap for 10 trials a day. On the last day all the animals jumped over the 15-cm gap. A medium-gray stimulus card was used during this period.

Phase 1. Discrimination training (Original learning). The animals were concurrently trained to criterion in Phase 1 training for 12 trials a day, 6 on each of two discrimination tasks: triangle versus circle, and vertical versus horizontal stripes. The animals were given three rewarded trials and three nonrewarded trials on each discrimination task per day. The criterion in Phase 1 training was that the median of the running times on the rewarded trials was shorter than the shortest running time on the nonrewarded trials for each task for 2 successive days. The positive and negative stimuli were counterbalanced. The order of the trials with the two tasks followed four predetermined random sequences. The order of rewarded and nonrewarded trials also followed four predetermined random sequences. On rewarded trials the animals were given two 45-mg milk pellets in the goal box, whereas on nonrewarded trials they were retained for 60 sec in the runway after the guillotine door was opened. Intertrial intervals ranged from 4 to 8 min.

Half of the animals received the same training for an additional 20 days after reaching the criterion in Phase 1 training (Group OT). The remaining animals received no further training on the original discrimination tasks once they had reached the criterion (Group NOT). Before the beginning of Phase 2 shift learning, the animals were given pretraining to habituate the apparatus (Y maze) for 10 trials on the day when they achieved the original learning criterion or accomplished overtraining. Medium-gray cards were used during this pretraining

Phase 2. Shift learning. After completing Phase 1 training, the animals were divided into two subgroups: C and IC matched with respect to the number of days to criterion. They were then trained under a given conditional successive discrimination condition in the Y maze until they achieved a criterion in Phase 2 shift. The criterion was 11 correct trials out of a possible 12 for each discrimination over 2 successive days combined. It is defined as an error response that the animals touch and push an incorrect stimulus and that they put their forelegs on the terrace on the front of the incorrect stimulus card. A self-correction training method was used in which, if the animals made an error, they were allowed to return the choice point and select a correct stimulus. Group C was run under the consistent training condition in which the animals were trained to choose the same goal box when the original positive stimulus or the negative one was presented on the entrance of each goal box. That is, they were trained to choose the right goal box when the original positive stimulus of each task was presented on the entrance of each goal box, and to choose the left goal box when the original negative stimulus of each task was presented, and vice versa. In contrast, Group IC was run under the inconsistent training condition in which the animals were trained to choose the different goal box in presenting the one original positive stimulus from that in presenting the other positive one of the two tasks as well as in the original negative stimuli. That is, they were trained to choose the right goal box when the positive stimulus of the vertical-horizontal stripe task was presented and to choose the left goat box when the positive one of the triangle-circle task was presented. Correspondingly, they were trained to choose the right goal box if the negative stimulus of the vertical-horizontal stripe task was presented and to choose the left goal box if the negative stimulus of the triangle-circle task was presented. The order of the trials with the two tasks followed four predetermined random sequences. Intertrial interval ranged from 4 to 8 min. The animals were given two 45-mg milk pellets in the goal box when they made the correct response.

Results

The group mean days-to-criterion on each discrimination task in Phase 1 training for each group are shown in Table 1. There was no indication of a difference among the four groups in the rate at which they learned in Phase 1 training, and this observation was supported by statistical analysis. An ANOVA using group (C vs. IC), degree of overtraining (NOT vs. OT), and task (H-V vs. C-T) revealed no significant main effects and no significant interactions (all Fs [less than] 1). The standard deviations of Table 1 appeared to be larger than the differences between the means. This was caused by equalizing the total number of days to criterion in the original learning for each group.

The results for each group in Phase 2 shift learning are illustrated in Figure 2. An ANOVA using group (C vs. IC) and degree of overtraining (NOT vs. OT) was performed on the number of days to criterion. The analysis revealed a significant effects of group [F(1, 20) = 5.44, p [less than] .03] and overtraining [F(1, 20) = 4.98, p [less than] .05]. That is, Group C learned Phase 2 shift more rapidly than Group IC after overtraining [t(10) = 2.94, p [less than] .05], but not after criterion training. Overtraining facilitated Phase 2 shift in Group C [t(10) = 3.54, p [less than] .01], whereas it did not facilitate Phase 2 shift in Group IC.
Table 1

Means and Standard Deviations of Number of Days to Criterion in
Phase 1 Training

 H-V T-Cir
Group/Task
 M SD M SD

C-NOT 12.33 5.12 13.17 4.95
C-OT 12.83 5.27 13.00 5.35
IC-NOT 13.67 4.75 12.67 5.41
IC-OT 12.50 6.21 12.67 5.59

Note. H = horizontal stripe stimulus, V = vertical stripe stimulus,
T = triangle stimulus, Cir = circle stimulus.


Discussion

Group C learned conditional successive discriminations in Phase 2 shift faster than Group IC did after overtraining, but not after criterion training. Overtraining facilitated Phase 2 shift learning in Group C, but not in Group IC. These findings suggest that animals form stimulus classes between the discriminative stimuli on a basis of the same response during overtraining. These findings agree with the cue-associations theory advocated by Nakagawa (1978, 1986, 1992). According to the cue-associations theory, animals learn a connection between a positive stimulus and an approach response as well as a connection between a negative stimulus and an avoidance response for each discrimination task in the original training. But during overtraining they also form associations between the discriminative stimuli with the same response assignment. Thus, animals learn to associate the two positive stimuli as well as to associate the two negative stimuli of the two independent discriminations. The associations between the two positive stimuli and between the two negative stimuli are called "cue associations." These cue associations produce an acquired equivalence effect, whereby stimuli associated with the same consequence show enhanced generalization; that is, cue associations mediate the transfer of appropriate responding from one positive (or negative) stimulus to the other positive (or negative) stimulus in reversal learning (Nakagawa, 1992, p. 52).

Alternatively, it is conceivable that stimulus classes formation in rats may be caused by reinforcement concordance, say of reinforcement or nonreinforcement, because rats are likely to form stimulus classes on a basis of the same consequence in a successive discrimination situation.

Experiment 2

The findings in Experiment 1 indicate that animals form stimulus classes between the discriminative stimuli on a basis of not only reinforcement concordance but also the same response (i.e., shared common response) during overtraining. To demonstrate that animals form stimulus classes on a basis of the same response, it is necessary to investigate directly whether or not animals form stimulus associations between the discriminative stimuli with the same response assignment (approaching response or avoidance response, for example). Thus, the present experiment was conducted to demonstrate that stimulus classes formation in rats was caused by the same response following the same consequence. In order to determine this, the animals were trained in two concurrent discriminations (A+B-, C+D- for example) to criterion or overtraining in Phase 1 training with a Y maze. Then they were given either whole reversal shift (both stimulus pairs reversed, from A+B-, C+D- to A-B+, C-D+ for example) or partial reversal shift (only one pair reversed, from A+B-, C+D- to A+B-, C-D+ for example) in Phase 2 shift with a straight runway alley in which two stimuli were presented on each trial. If animals formed stimulus classes between the discriminative stimuli on a basis of the same response during overtraining, they should then learn whole reversal more rapidly than partial reversal after overtraining, but not for after criterion training.

Method

Subjects. Thirty-two experimentally naive mate Sprague-Dawley rats were used. They were about 160 days old with an initial average body weight of 348 g. All details of feeding schedule and handling were the same as in Experiment 1. The animals were maintained on a 14:10-hr light:dark cycle with light off at 11:00 p.m.

Apparatus. Both a modified Y maze which was the same as in Experiment 1 and a double straight runway shown in Figure 3 were used. The double straight runway consisted of a runway with a startbox, a goal box, and two gaps over which the animals had to jump. A guillotine door was located at the front of a startbox. The first gap was located 40 cm in front of the startbox. The second gap was located farther away, at 85 cm. Pieces of cardboard were placed 5 cm in front of the first gap (first stimulus presentation position: SPP-I) and at the entrance of the goal box (second stimulus presentation position: SPP-II) which served as discriminanda. The apparatus was painted medium-gray inside and lit through the experiment by two 10-W fluorescent lamps suspended 45 cm above the top of runway. One fluorescent lamp was suspended between the guillotine door and SPP-I, and the other was suspended between SPP-I and SPP-II. Separate starting time and running time were obtained on each trial by means of two electrical digital timers. When the experimenter raised the guillotine door, a microswitch activated the first timer. When the animals interrupted a photobeam gate 3 cm down the runway, this timer stopped. At the same time the second timer started. When the animals interrupted the second photobeam gate 55 cm farther down the runway, the second timer stopped. A running time is the time between the interrupting of the first and the second photobeam gates. That is, the distance for running time was 55 cm.

Stimuli. Stimulus cards were 12-cm squares of cardboard. In the Y maze, each square was presented at the entrance of each goal box and served as an entrance door. They were arranged so that the card serving as the correct door could be pushed down easily, thus permitting the animals to gain entrance into the goal box; the card denoting the incorrect door was locked. By contrast, in the straight runway, cards were presented at SPP-I and SPP-II in which the cards served as an entrance door. They were arranged so that on rewarded trials both cards presented at SPP-I and SPP-II, in which the cards served as the correct door, could be pushed down easily, thus permitting the animals to gain entrance into the goal box. In contrast, on nonrewarded trials the card presented at SPP-I could be pushed down easily but the one denoting the incorrect door at SPP-II was locked. For a white-black discrimination a white card and a black one were used. Vertically striped and horizontally striped cards were used for a vertical-horizontal stripe discrimination which were the same as in Experiment 1.

Procedure. The animals were given pretraining for 8 days prior to the beginning of discrimination training in the Y maze. That is, on Day 1 the rats were allowed to explore the apparatus for two periods of 10 and 7 min. From Day 2 to Day 4, the animals were trained to push down a stimulus card and enter the goal box to obtain food for 10 daily trials. The gap between the arms and goal boxes was not present for this stage of the experiment. From Day 5 to Day 8 rats were trained to jump over the gap for 10 trials a day. On the last day all the animals jumped over the 15-cm gap. They were given the same number of trials on each arm during this pretraining. That is, one of the arms was blocked off during pretraining. A trial in this experiment is defined as a response-stimulus sequence when the animals start from the starting box after opening the guillotine door, run down in the runway, push down a correct stimulus, and enter the goal box to obtain food. Medium gray stimulus cards were used during this period.

Phase 1. Discrimination training (Original learning). The animals were concurrently trained to a criterion in the original learning for 12 trials a day on two discrimination tasks: white versus black and vertical versus horizontal stripe. The criterion was 11 correct trials out of a possible 12 for each discrimination over 2 successive days combined. At the end of a trial, the animals were removed from the goal box by the experimenter and placed in an individual waiting cage. A noncorrection training method was used. The positive and negative stimuli were counterbalanced. The order of trials with the two tasks followed four predetermined random sequences. The position of a positive stimulus also followed four predetermined random sequences. The animals were given two 45-mg milk pellets when they made a correct response. Intertrial intervals ranged from 4 to 8 min.

After reaching the criterion, half of the animals were given overtraining on the original discrimination tasks for 20 days (Group OT). The remaining animals received no further training on these tasks once they had reached the criterion (Group NOT).

Before the beginning of the reversal learning in the double straight runway alley, the animals were given pretraining to habituate the apparatus for 10 trials on the day when they achieved the criterion or accomplished overtraining. Medium-gray stimulus cards were used during this pretraining in the double straight runway.

Phase 2. Reversal learning. After completing Phase 1 training, the animals of each group were divided into two subgroups: W and P, matched with respect to the number of days to criterion. They were then trained under a given reversal condition until they achieved a criterion in Phase 2 reversal. The criterion was that the median of the running times on rewarded trials was shorter than the shortest running time on nonrewarded trials for each task for two successive days. Group W was run under a whole reversal condition, in which both discriminations were reversed, but the relations of the stimuli associated with the same response assignment in Phase 1 training were kept intact. That is, the positive stimulus of the vertical-horizontal stripe task in the original training was presented at SPP-I and the positive stimulus of the white-black was presented at SPP-II on nonrewarded trials, and vice versa. On rewarded trials, the negative stimulus of vertical-horizontal stripe was presented at SPP-I and the negative of the white-black was presented at SPP-II, and vice versa. By contrast, Group P was run under a partial reversal condition, in which only the vertical-horizontal stripe task was reversed, but the other task continued the same training as one in Phase 1 training. Thus, in this partial reversal, the relations of the stimulus associated with the same outcome in Phase 1 training were not kept intact. That is, the positive stimulus of the vertical-horizontal stripe in Phase 1 training was presented at SPP-I and the negative of the white-black was presented at SPP-II on nonrewarded trials, and vice versa. On rewarded trials the negative stimulus of the vertical-horizontal stripe was presented at SPP-I and the positive of the white-black was presented at $PP-II and vice versa. The position of the two tasks followed four predetermined random sequences. The order of rewarded and nonrewarded trials also followed four predetermined random sequences. On rewarded trials the animals were given two 45-mg milk pellets in the goal box. By contrast, on nonrewarded trials they were retained for 45 sec in the runway either after they touched the incorrect stimulus presented at SPP-II, or after the guillotine door was opened in the case of occurring nonresponding. The animals were allowed to run down or up freely in the section between the guillotine door and SPP-II for 45 sec on non rewarded trials. Intertrial interval ranged from 4 to 8 min.

Results

The group mean days-to-criterion on each discrimination task in Phase 1 training for each group are shown in Table 2. There was no indication of a difference among four groups in the rate at which they learned in Phase 1 training, and this observation was supported by statistical analysis. An ANOVA using group (W vs. P) and degree of overtraining (NOT vs. OT) and task (W-B vs. H-V) revealed that neither main effects nor interactions were significant (all Fs [less than] 1). The standard deviations of Table 2 appeared to be larger than the difference between the means. This was caused by equalizing the total number of days to criterion in the original learning for each group. The percentage of errors during overtraining was 5.1%.
Table 2

Means and Standard Deviations of Number of Days to Criterion in
Phase 1 Training

 B-W H-V
Group/Task
 M SD M SD

W/OT 13.88 5.30 15.63 5.85
W/NOT 14.38 5.18 15.25 4.98
P/OT 14.63 5.53 15.50 5.37
P/NOT 15.88 5.54 15.63 5.26

Note. B = black stimulus, W = white stimulus, H = horizontal stripe
stimulus, V = vertical stripe stimulus.


The results for each group in Phase 2 reversal are illustrated in Figure 4. An ANOVA using group (W vs. P) and overtraining (NOT vs. OT) was performed on the number of days to criterion on the common reversal task (vertical-horizontal stripe), which revealed the statistically significant effect of group [F(1, 28) = 14.35, p [less than] .01] and a statistically significant degree of overtraining x group interaction [F(1, 28) = 71.61, p [less than] .001]. Overtraining significantly facilitated whole reversal IF(1, 28) = 23.75, p [less than] .001], whereas it delayed partial reversal [F(1, 28) = 50.31, p [less than] .001]. The animals learned whole reversal more rapidly than partial reversal after overtraining [F(1, 28) = 75.05, p [less than] .001), whereas they learned partial reversal more rapidly than whole reversal after criterion training [F(1, 28) = 10.91, p [less than] .01].

Overtraining significantly facilitated reversal of the white-black task in Group W [mean days to criterion was 27.13 (SD = 4.37) for overtrained rats and 18.13 (SD = 2.12) for nonovertrained ones] [t(14) = 4.89, p[less than] .01].

In order to examine the degree of interaction of performance on the nonreversal task and the reversal one in Group P, a special criterion was devised: If the shortest of running times on nonrewarded trials became shorter than the median of the running times on rewarded trials for the nonreversal task, say of the white-black task, the day was determined as a retention loss day. The number of these days was counted for each rat. The number of retention loss days for each group is as follows: 23.33 (SD = 10.40) for overtrained rats and 9.67 (SD = 3.68) for nonovertrained rats. Overtraining significantly increased the retention loss days on the nonreversal task [F(1, 14) = 6.54, p [less than] .01].

Discussion

This experiment essentially replicated the pattern of results seen in the first study: The same response (i.e., shared common response) effect was observed in that Group W learned Phase 2 reversal faster than Group P after overtraining, but not after criterion training, and overtraining facilitated Phase 2 reversal in Group W, but not in Group P. These findings are strong evidence for the animals to form associations between the discriminative stimuli on a basis of the same response during overtraining

The basic whole-partial reversal effect was replicated in this experiment in that Group W reversed more rapidly than Group P after overtraining, but not after criterion training. Positive ORE (Overtraining Reversal Effect) occurred in Group W, whereas negative ORE occurred in Group R These findings are in line with the cue-associations theory (Nakagawa, 1978, 1986, 1992). That is, according to the cue-associations theory, as a result of cue associations, the reversal of the one discrimination after overtraining in the whole condition should exert a synergistic influence upon the reversal of the other discrimination. Each reinforcement of the new positive stimulus in one discrimination should not only enhance the strength of the approach response to this stimulus but also augment the same response to the new positive stimulus in the other discrimination via the cue association between the positive stimuli formed during overtraining. Correspondingly, the consequences of nonreinforcement of the new negative stimulus should also transfer between discriminations. Thus, the cue-associations theory predicts a positive ORE in whole condition because in the criterion training there are no cue associations present to mediate the synergistic interaction between discriminations.

A parallel explanation can be given for the retarded reversal observed in the partial condition after overtraining. In the presence of cue associations, continued training with the non reversed discrimination during the reversal stage will lead to interference with the development of an approach response to the new positive stimulus and of an avoidance response to the new negative stimulus of reversed discrimination. Consequently, Group P should exhibit a negative ORE. This is supported by the retention loss day data.

Group P reversed more rapidly than Group W after criterion training. This finding indicates that the four stimulus-outcome associations in the original training is independent. Group P had only two associations to relearn, whereas Group W had four associations to relearn. Thus, one would have expected that Group P would have reversed faster than Group W.

The aim of using the double straight runway as an apparatus in Phase 2 reversal is to examine directly whether or not the animals form associations between the discriminative stimuli on a basis of the same response. In this apparatus, the animals experienced two stimuli on each trial; one was presented at SPP-I and the other was presented at SPP-II, before they received a reward or nonreward in the goal box on each trial. That is, their responding to the stimulus presented at SPP-II was directly reinforced, whereas their responding to the stimulus presented at SPP-I was secondarily reinforced. Therefore, to accomplish reversal shift in Phase 2, the animals were required to relearn the stimulus (presented at SPP-I) - the stimulus (presented at SPP-II) - food or no food associations. If the animals formed stimulus associations between the discriminative stimuli on a basis of the same response after overtraining in Phase 1 training, for the animals of Group W-OT, the two stimuli presented at SPP-I and SPP-II on both rewarded trials and nonrewarded ones should have already been associated with each other on the basis of the same response. Thus, they were required to connect newly these two stimuli associations with food or no food. In contrast, the animals of Group W-NOT should be required to relearn associations between these two stimuli presented at both SPP-I and SPP-II because they did not form stimulus classes between the discriminative stimuli in Phase I training and then they had to associate newly these discriminative stimuli with food or no food. By contrast, for the animals of Group P-OT, two stimuli presented at both SPP-I and SPP-II on each trial of rewarded and nonrewarded trials have belonged in different stimulus classes established during overtraining. And so the animals of Group P-OT were required to approach both the one original positive stimulus at SPP-I and the one negative stimulus at SPP-II, vice versa, on rewarded trials. On nonrewarded trials, they were required to avoid both the other original positive stimulus at SPP-I and the other original negative stimulus at SPP-II, vice versa. Thus, they were required to dissociate stimulus classes established during overtraining and to relearn newly stimulus associations between these two stimuli presented at both SPP-I and SPP-II on a basis of either go response or no-go response following reinforcement or nonreinforcement. On balance, the animals of Group P-NOT should relearn newly associations between the two stimuli presented at both SPP-I and SPP-II because they did not form the stimulus classes between the discriminative stimuli and then they did not need to dissociate the stimulus classes. Therefore, the findings of this experiment that Group W reversed more rapidly than Group P after overtraining, a positive ORE in Group W and a negative ORE in Group P demonstrate that the animals form stimulus classes on the basis of the same response during overtraining. Indeed, if the animals would form stimulus classes between the discriminative stimuli in criterion training, Group W should then reverse faster than Group P after criterion training. But this result was not observed.

General Discussion

Two experiments have examined whether or not the animals form stimulus classes on a basis of the same response (i.e., shared common response) during overtraining. In Experiment 1, the animals were concurrently trained on two discrimination tasks: white versus black and vertical versus horizontal stripe in a straight runway. Then they were trained on conditional successive discriminations in a Y maze. The animals of Group C learned the conditional successive discriminations more rapidly than Group IC after overtraining, but not after criterion training. Overtraining significantly facilitated the conditional successive discriminations in Group C, but not in Group IC. In Experiment 2, the animals were concurrently trained on two discrimination tasks: white versus black and vertical versus horizontal stripe in the Y maze. They were then trained on either whole reversal tasks (Group W) or partial reversal one (Group P) in a double straight runway. The animals of Group W mastered their reversal more rapidly than Group P after overtraining, but not after criterion training. Overtraining significantly facilitated the reversal shift in Group W, but not in Group P. In addition, the animals of Group P-OT showed, greater retention loss of the original white-black discrimination than their counterparts that were not overtrained, Group P-NOT. These results of two experiments make it clear that the animals form stimulus classes between the discriminative stimuli on a basis of the same response during overtraining in the conditional discrimination learnings; this is the novel contribution of the present study.

There was no significant difference in the rate of the shift learning between Group C and Group IC after criterion training, whereas Group C learned the shift problems more rapidly than Group IC after overtraining in Experiment 1. Group W reversed in fewer days than Group P after overtraining, whereas Group P reversed in fewer days than Group W after criterion training in Experiment 2. These findings demonstrate that the animals establish stimulus classes between stimuli on a basis of the same response during overtraining in the conditional discrimination learning but not in criterion training. These findings are in line with the cue-associations theory (Nakagawa, 1978, 1986, 1992). That is, these findings make it clear that the animals need two stages to establish stimulus classes between the discriminative stimuli on a basis of the same response in the conditional concurrent discrimination learning: The first stage is that the animals learn a connection between a positive stimulus and an approach response followed by food as well as a connection between a negative stimulus and an avoidance response followed by no food for each discrimination task on reaching the criterion in Phase 1 discrimination training, and the second one is that they form associations between the discriminative stimuli on a basis of the same response following the same consequence during overtraining.

The fact that Group W learned their reversal more rapidly than Group P after overtraining but not after criterion training in Experiment 2 is in line with findings of Nakagawa (1978, 1986, 1992) in rats and Delius et al. (1995) in pigeons. This fact provides strong evidence that rats form stimulus classes between the discriminative stimuli with the same response assignment in conditional concurrent discrimination learning.

A stimulus association view based on reinforcement concordance in the formation mechanism of stimulus classes has two positions: One is stimulus associations based on downright reinforcement concordance, say of reinforcement or nonreinforcement (Edwards et al., 1982; Hall et al., 1993; Urcuioli et al., 1989; Zentall, Steirn, et al., 1991; Zentall, Sherburne, et al., 1992). The other is stimulus associations on a basis of the same response following the same consequence (Nakagawa, 1978, 1986, 1992, 1993, 1998; Delius et al., 1995). However, stimulus association mechanisms postulated by Nakagawa (1978, 1986, 1992, 1993, 1998) and Delius et al. (1995) are not necessarily in conflict with stimulus association process based on downright reinforcement concordance (Edwards et al., 1982; Hall et al., 1993; Urcuioli et al., 1989; Zentall, Sherburne, et al., 1992, Zentall, Steirn, et al., 1991). Conversely, it seems likely that in nature both principles may often act in consonance.

The results of the present experiments reported here provide strong evidence that the same response following the same consequence during overtraining is a factor affecting stimulus classes formation in concurrent discriminations in rats. Thus, the present results may contribute to understanding a mechanism of stimulus classes formation in concurrent discriminations in rats.

References

BHATT, R. S., & WASSERMAN, E. A. (1989). Secondary generalization and categorization in pigeons. The Journal of the Experimental Analysis of Behavior, 52, 213-224.

DELIUS, J. D., AMELING, M., LEA, S. E., & STADDON, J. E. R. (1995). Reinforcement concordance induces and maintains stimulus associations in pigeons. The Psychological Record, 45, 283-297.

DUBE, W. V., CALLAHAN, T. D., & MCILVANE, W. J. (1993). Serial reversal of concurrent auditory discriminations in rats. The Psychological Record, 43, 429-440.

EDWARDS, C. A., JAGIELO, J. A., ZENTALL, T. R., & HOGAN, D. E. (1982). Acquired equivalence and distinctiveness in matching to sample by pigeons. Journal of Experimental Psychology: Animal Behavior Processes, 8, 244-259.

FERSEN, L., von, & LEA, S. E.G. (1990). Category discrimination with polymorphous feature. The Journal of the Experimental Analysis of Behavior, 54, 69-84.

HALL, G., RAY, E., & BONARDI, C. (1993). Acquired equivalence between cues trained with a common antecedent. Journal of Experimental Psychology: Animal Behavior Processes, 19, 391-399.

NAKAGAWA, E. (1978). The effect of overtraining on discrimination learning in white rats (in Japanese). Japanese Journal of Psychology, 49, 70-77.

NAKAGAWA, E. (1986). Overtraining, extinction and shift learning in a concurrent discrimination in rats. Quarterly Journal of Experimental Psychology, 38B, 213-226.

NAKAGAWA, E. (1992). Effects of overtraining on reversal learning by rats in concurrent and single discriminations. Quarterly Journal of Experimental Psychology, 44B, 37-56.

NAKAGAWA, E. (1993). Relational rule learning in the rat. Psychobiology, 21, 293-298.

NAKAGAWA, E. (1998). Stimulus classes formation in concurrent discriminations in rats as a function of overtraining. The Psychological Record, 48, 537-560.

SPRADLIN, J. E., COTTER, V. W., & BAXLEY., N. (1973). Establishing a conditional discrimination without direct training: A study of transfer with retarded adolescents. America Journal of Mental Deficiency, 77, 556-566.

URCUIOLI, P. J., ZENTALL, T. R., JACKSON-SMITH, P., & STEIRN, J. N. (1989). Evidence for common coding in many-to-one matching: Retention, intertrial interference, and transfer. Journal of Experimental Psychology: Animal Behavior Processes, 15, 264-273.

VAUGHAN, W. (1988). Formation of equivalence sets in pigeons. Journal of Experimental Psychology: Animal Behavior Processes, 14, 36-42.

VAUGHAN, W., & HERRNSTEIN, R. J. (1987). Choosing among natural stimuli. The Journal of the Experimental Analysis of Behavior, 47, 5-16.

ZENTALL, T. R., SHERBURNE, L. M., STEIRN, J. N., RANDALL, C. K., ROPER, K. L., & URCUIOLI, P. J. (1992). Common coding in pigeons: Partial versus total reversals of one-to-many conditional discriminations. Animal Learning & Behavior, 20, 373-381.

ZENTALL, T. R., STEIRN, J. N., SHERBURNE, L. M., & URCUIOLI, P. J. (1991). Common coding in pigeons assessed through partial versus total reversals of many-to-one conditional and simple discriminations. Journal of Experimental Psychology: Animal Behavior Processes, 17, 194-201.

2.41
COPYRIGHT 1999 The Psychological Record
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1999 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Nakagawa, Esho
Publication:The Psychological Record
Date:Jan 1, 1999
Words:7630
Previous Article:Errors and response latencies as a function of nodal distance in 5-member equivalence classes.
Next Article:Memory distortion: can accurate memory be preserved?
Topics:


Related Articles
The effects of equivalence class structure on test performances.
The search for stimulus equivalence in nonverbal organisms.
Emergent relations in the formation of stimulus classes by pigeons.
The effects of terminal-link stimulus arrangements on preference in concurrent chains.
The quick development of equivalence classes in a paper-and-pencil format through written instructions.
Stimulus equivalence and the blocking effect.
Mechanism of stimulus classes formation in concurrent discriminations in rats.
EFFECTS OF OVERTRAINING ON REVERSAL AND NONREVERSAL LEARNING ON CONCURRENT DISCRIMINATIONS IN RATS.
TRANSFER OF LEARNING BETWEEN MATCHING (OR NON-MATCHING)-TO-SAMPLE AND SAME-DIFFERENT DISCRIMINATIONS IN RATS.
Overtraining, extinction, and shift learning in matching-to-sample discriminations in rats.

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters