- Email: [email protected]
Different effects of amphetamine on reinforced variations versus repetitions in spontaneously hypertensive rats (SHR)
Physiology & Behavior, Vol. 56, No. 5, pp. 939-944, 1994 Copyright © 1994 ElsevierScienceLtd Printed in the USA. All rights reserved 0031-9384/94 $6.00 + .00
Pergamon 0031-9384(94)00160-X
Different Effects of Amphetamine on Reinforced Variations Versus Repetitions in Spontaneously Hypertensive Rats (SHR) DEBORAH
M. MOOK
AND
ALLEN
NEURINGER I
Department o f Psychology, Reed College, Portland, OR 97202 R e c e i v e d 23 A u g u s t 1993
MOOK, D. M. AND A. NEURINGER. Different effects of amphetamine on reinforced variations versus repetitions in spontaneously hypertensive rats (SHR). PHYSIOL BEHAV $6(5) 939-944, 1994.--The spontaneously hypertensive rat (SHR) may serve as an animal model of human attention deficit hyperactivity disorder (ADHD). We compared performances of SHRs and Wystar-Kyoto normotensive controls rats (WKY) in two experiments. When rewarded for varying sequences of responses across two manipulanda, the SHRs were more likely to vary than the WKYs. On the other hand, when rewarded for repetitions of a small number of sequences, the WKYs were more likely to learn to repeat. Both of these results confirm previous findings. Injecting 0.75 mg/kg d-amphetamine facilitated learning by SHRs to repeat the required sequences, with amphetamine-injected SHRs learning as rapidly as saline-injected, control WKYs. On the other hand, amphetamine tended to increase variability in both strains when high levels of variations were required for reward, and to decrease it in both strains when low levels of variability were required. Thus, amphetamine may have different effects on reinforced repetitions vs. reinforced variations. SHR rats ADHD Behavioral variability Amphetamine Reinforcement Operant conditioning Selection
THE spontaneously hypertensive rat (SHR), originally studied because of its high blood pressure (24), may also serve as an animal model for another human disorder, attention deficit hyperactivity disorder, or A D H D (29,32). There are two main sources of support for this hypothesis. First, SHRs are more active than their normotensive W i s t a r - K y o t o (WKY) progenitors, typically used as controls (6,9,10,16,31,32). Children with A D H D have also been shown to be more active than controls (26,27). Second, SHRs take longer to learn some tasks than WKYs and learn to lower levels of proficiency (14,23). A D H D children may also show such learning deficits (1,3,5,26). Thus, high levels of activity and learning disabilities have been reported both for children diagnosed with A D H D and SHR rats ( 1,3,5,6, 8,9,12,13,18,26,27,30- 32). SHRs do not always show learning deficits, however. For example, they avoid or escape aversive stimuli more readily than WKYs (7,11). This finding may partly be accounted for by the fact that aversive stimuli elicit different responses in the two s t r a i n s - - S H R s tend to move and W K Y s to freeze (2). Sensitivity to painful stimulation may also differ in the two strains [(15), see, however, (l 0)]. Even under some positively reinforcing conditions, SHRs learn more rapidly than W K Y s (e.g., in the H e b b -
Hyperactivity
Response sequences
Memory
William's maze) (28). Thus, a general learning deficit does not appear to characterize SHR performance. Two studies may help to explain why SHRs sometimes perform less successfully than WKYs, while at other times they are more successful. The original clue was provided by Low, Whitehorn, and Hendley (14). When reinforcement depended on rats learning to go to a single arm of a T-maze, WKYs were more proficient than SHRs; but when reinforcement depended on rats varying their arm entries, SHRs excelled. Mook, Jeffrey, and Neuringer (16) extended the Low et al. (14) findings with two different procedures. (a) In a 12-arm radial maze task, where variations of arm entries are required, SHRs varied their arm choices more, making fewer repetition errors, than WKYs. On the other hand, when the requirement for response variability was combined with a requirement to repeat selected arms in the radial maze, WKYs more rapidly learned to repeat the required arms. (b) Similar results were obtained in an operant chamber: when reinforcement depended on variable sequences of left and fight responses across two levers, the SHRs responded more variably than WKYs, but when sequences had to be repeated, the WKYs better learned to repeat. Thus, in both the Low et al. and Mook et al. studies, strain of rats interacted with reinforcement contin-
To whom requests for reprints should be addressed. 939
940
MOOK AND NEURIN(;ER
gencies in determining whether SHRs showed a learning deficit or advantage. When required to vary, SHRs had an advantage over WKYs; when required to repeat, WKYs had the advantage. ADHD in humans is commonly treated with stimulant medication (e.g., Dexedrine, Ritalin, or Cylert), which causes diagnosed subjects to behave in a manner approximating that of normal controls (1,8,12,26,27). Several studies have explored the effects of stimulant medication on SHR's locomotor responses and response rates (19,29-32). The present study extends that research by asking whether amphetamine has different effects when response variations vs. repetitions are required. The general question is whether the effects of amphetamine on SHRs vs. WKYs depend upon reinforcement contingencies. GENERAL METHOD
Subjects Fourteen male SHR and 14 male WKY, obtained from Harlan Sprague-Dawley, Inc. (Indianapolis, IN), were housed in pairs in 40.5 x 24 × 19 cm cages under a 12-h light/dark cycle, and maintained on a 21-h food-deprivation schedule, with water available freely except during the experimental hour. The same subjects, experimentally naive at the beginning, were used in all experiments.
Apparatus Nine identical modified Gerbrands operant chambers were 27 cm deep, 28 cm long, and 30 cm high. The ceiling, back wall, and front wall (which opened on hinges) were constructed of clear Plexiglas, and the two side walls were constructed of aluminum. The floor consisted of wire bars 1 cm apart. One side wall contained three pigeon response keys that were 3 cm in diameter, 17.5 cm from the ceiling, and 9.5 cm from the floor. The keys were 9 cm apart from center to center, and could be illuminated by white 24-V bulbs. An audio speaker was located behind the wall containing the keys. Only two keys were act i v e - t h e middle key and the right key. On the opposite wall were two Gerbrands response levers that were not used in this experiment. Located between the two levers was a pellet tray, 3 cm from the floor, into which 45-mg Noyes pellets could be released. Each chamber was housed in a sound- and light-attenuating outer cubicle made of wood and connected to an Apple Macintosh computer through a Metaresearch Benchtop interface. These cubicles were equipped with one-way mirrors through which subjects could be observed during sessions.
trial consisted of lour pushes on left (L) or right (R) keys. If the current sequence differed from each of the previous R~ur sequences (a lag 4 variability contingency), then a lk)od pellet was presented. Thus, for example, if the prior sequences had been LRRR, RRRR, RLLR, and RRRR, then an LLRL sequence would end with reinforcement, because it differed flom each of the previous four, whereas an RLLR sequence would end with a brief time-out, because it repeated one of the previous four sequences. METHOD
Procedure Subjects were first autoshaped to press each of the two keys for food reinforcement, and then placed on a fixed ratio (FR) schedule, the value of the ratio increasing to FR4 over the course of four sessions. Four responses summed across the keys were required for reinforcement. The variability schedule was then introduced, in which four presses on left and right keys constituted a trial, and the current sequence was required to differ from each of the sequences in the preceding four trials for the animal to be reinforced, a lag 4, variability contingency (17,25). As in previous studies, the four-trial (or lag) window moved at the end of each trial, so that the current sequence always had to differ from the just-prior four sequences for reinforcement to be delivered. When the variability contingency had been met, a food pellet was presented, preceded by a 0.5-s end-of-trial cue, namely darkening of the chamber and a 200 Hz tone alternating with a 1000 Hz tone, creating a warbling sound. If the sequence in the current trial repeated any one or more of the previous four sequences, then a 3-s time-out was imposed, consisting of the same dark chamber but with no tone and no reinforcer. After time-out or reinforcement, the two keys were reilluminated and a new trial began. Sessions terminated after 100 trials or 0.5 h, whichever occurred first. Ten sessions were provided, 5 days per week.
Data Analysis Percent vary, the main dependent variable, was calculated by dividing the number of sequences in a session that differed from each of the previous four sequences (i.e., number of reinforced sequences) by the total number of sequences (or total trials)in the session, the result multiplied by 100 to obtain a percentage. RESULTS AND DISCUSSION Figure 1 shows group average percent vary across 10 sessions. A 2 x 2 × 10 (strain X drug x sessions) ANOVA showed that
Drugs and Dosage d-Amphetamine sulfate was dissolved in 0.9% saline and injected IP in a dose of 0.75 mg/kg by body weight. Fifteen minutes prior to each session, all animals were weighed and administered with either the drug solution or with an equal volume of 0.9% saline according to group assignment. Assignment to amphetamine vs. saline conditions was random, with seven SHRs and seven WKYs receiving saline and seven SHRs and seven WKYs receiving amphetamine. Thus, the four experimental groups were: SHR-amphetamine (SHR-AMPH), SHR-saline (SHRSAL), WKY-amphetamine (WKY-AMPH), and WKY-saline WKY-SAL). EXPERIMENT 1: OPERANT VARIABILITY Variable sequences of left and right responses were reinforced in an operant chamber, a task similar to that previously used with rats and pigeons [e.g., (16,17,20,22,25)]. In the present case, a
8070-
tr ,~
50
4O 30
WKY..SAL
---"0--
!
!
!
2
4
6
-
|
8
. "'"
I
10
SESSIONS
FIG. 1. Percentage of reinforcements for SHRs and WKYs given amphetamine and saline under the variability contingency during each of 10 sessions.
SHR AND AMPHETAMINE
there was a significant strain effect, F(1, 24) = 8.569, p < 0.01, with SHRs responding more variably than WKYs; a weak drug effect, F(1, 24) = 3.622, p < 0.10, with subjects given amphetamine responding more variably than those given saline; and a significant sessions effect, F(9, 216) = 3.954, p < 0.001, with variability increasing across sessions. None of the interactions reached significance. Two additional planned comparisons were performed to test (i) whether nondrugged SHRs responded more variably than nondrugged WKYs, a finding reported in previous studies, and (ii) whether amphetamine administered to SHRs would cause them to behave similarly to control, nondrugged WKYs, a comparison relevant to treatment of human ADHD with stimulant medication. We found that (i) saline-injected SHRs were indeed significantly more variable than similarly injected WKYs, F(1, 24) = 8.273, p < 0.01, results consistent with previous findings (14,16). However, (ii) amphetamine administered to SHRs did not cause them to perform like saline-injected WKYs, with the SHRAMPH group responding significantly more variably than the WKY-SAL group, F(1, 24) = 11.667, p < 0.01. This comparison therefore did not support the hypothesis that amphetamine causes SHRs, a proposed animal model of human ADHD, to behave like normal animals. The next experiment further tested these questions in a situation where both variations and repetitions were required. EXPERIMENT 2: OPERANT REPETITION AND VARIATION Experiment 2 employed the Mook et al. (16) procedure. Rats were required to select repeatedly 4 of 16 possible sequences, but could not repeat the same sequence emitted in the just-previous trial (i.e. a lag 1 variability contingency was imposed within the reinforcable subset). The main goal was to test the Experiment 1 findings that amphetamine affected SHRs and WKYs similarly and that amphetamine-injected SHRs differed significantly from saline-injected control subjects.
941
lection was calculated by dividing the number of within-subset sequences by the total number of sequences, multiplied by 100. The numerator included both reinforced and nonreinforced (or repeated) entries into the designated subset. Percent vary was calculated by dividing the number of reinforced sequences by total number of emissions of the subset, multiplied by 100. Thus, a high value for percent selection indicated that the rat learned to emit the four potentially reinforced sequences and to avoid emitting the other sequences. A high value for percent vary indicated that many within-subset sequences were preceded by a different sequence, either a different within-subset sequence or any nonsubset sequence. RESULTS AND DISCUSSION
Selection Figure 2 (top) shows group average percent selection across the 12 sessions of Experiment 2a. The figure indicates that performances approached asymptote at about the ninth session. The data were therefore divided into two parts for ease of analysis: acquisition (nine sessions of learning) and asymptote (three final sessions). A 2 × 2 × 9 ANOVA (strain × drug × sessions) on percent selection (acquisition) showed a significant effect of strain, F(I, 23) = 7.553, p < 0.05, with WKYs selecting the correct subset more frequently than SHRs; a significant effect of drug, F(1, 23) = 5.008, p < 0.05, with amphetamine-injected rats selecting the correct subset more frequently than saline-injected; and a significant effect of sessions, F(8, 184) = 28.639, p < 0.0001, with performances improving across sessions. None of the interactions approached significance. The first of two additional planned com80 70 ¸
METHOD
Procedure As in Experiment 1, four responses to the two keys constituted a trial. However, in the present experiment, only four sequences could be reinforced with the other 12 sequences never reinforced. The potentially reinforced sequences in Experiment 2a, defined as the subset, were RRRR, RRRL, RRLR, and RRLL. A sequence in this subset was reinforced only if it differed from the previously emitted sequence (i.e., a lag 1 variability contingency). To optimize reinforcement, subjects therefore had to learn: (a) to emit only the four sequences in the subset, and (b) not to emit the same sequence twice in a row. Subjects were run for 12 sessions, generally five sessions per week, with each session terminating after 100 trials or 0.5 h, whichever occurred first. Experiment 2b repeated the procedure with one difference. The subset was changed to LLLL, LLLR, LLRL, and LLRR (i.e., the mirror-image sequences of those used in Experiment 2a). As will be seen, the results of Experiment 2a were of some interest, and Experiment 2b attempted to replicate 2a with sequences that were most comparable in difficulty but for which there would be minimal direct transfer [see (21)]. Experiment 2b followed immediately after 2a.
Data Analysis The two dependent measures were percent selection of the required subset and percent vary within the subset. Percent se-
5O LU
¢/)
40
~
3o 20
10 0
80 70 60 50 40
3o
,,O"
,'~ • ;/"
"~'O'''O'- .O"-
-
-
~-
=u,~_AuI SHR-AMPH
20 10
I
i
l
I
2
4
6
8
I0
1
I
12
SESSIONS FIG. 2. Percentage of trials in which SHRs and WKYs injected with amphetamine or saline emitted any of the four poteniaUy reinforced response sequences. The top graph shows Experiment 2a, the bottom shows Experiment 2b.
942
parisons showed that saline-injected WKYs selected the correct subset more frequently than saline-injected SHRs, F(1, 23) 5.656, p < 0.05, again replicating the previous findings. The second planned comparison showed that SHRs injected with amphetamine were statistically indistinguishable from control WKYs administered saline, F(l, 23) = 0.136, NS. The improvement due to amphetamine approached significance for SHRs (p < 0.06), but was not statistically significant for WKYs. Thus, unlike Experiment l, but consistent with human ADHD findings, amphetamine caused SHRs to approximate the performance of nondrugged control subjects. Similar results were obtained in Experiment 2b, as shown in Fig. 2 (bottom). There were significant strain, F(1,23) = 18.993, p < 0.001, drug, F(l, 23) -- 5.165,p < 0.05, and sessions effects, F(8, 184) = 54.464, p < 0.0001, and, in this case, a significant drug by strain interaction, F(1, 23) = 5.496, p < 0.05. As in Experiment 2a, the saline-injected WKYs learned more rapidly than the saline-injected SHRs, F(l, 23) = 23.397, p < 0.0001, and injecting the SHRs with amphetamine caused them to learn as rapidly as the control WKYs, F(1, 23) = 2.265, NS. Again as in Experiment 2a, amphetamine facilitated learning to repeat by the SHRs, F(1, 23) = I 1.02, p < 0.01, but not the WKYs, F(l, 23) = 0.002, NS. Thus, Experiment 2b essentially replicated the results of Experiment 2a with respect to repetitions of subsets of responses. Two 2 × 2 × 3 ANOVAs (strain × drug × sessions) on percent selection at asymptote (i.e., during the last three sessions of the experiments) yielded no significant differences. In each of the two experiments, the groups attained levels of performance that did not differ at asymptote. Variation
Figure 3 (top) shows percent vary within the subset in Experiment 2a. The data were again divided into acquisition and asymptote phases. A 2 X 2 x 9 ANOVA (strain × drug x sessions) of the percent vary (acquisition) data showed a weak strain effect, F(1, 23) = 3.017, p < 0.10, with SHRs behaving more variably than WKYs, no significant effect of drug, and a significant sessions effect, F(8, 184) = 12.449,p < 0.0001. The strain by sessions interaction was significant, F(8, 184) = 2.758, p < 0.01, with SHRs behaving more variably than WKYs during the initial sessions. The strain by drug by sessions interaction also attained significance, F(8, 184) = 2.024, p < 0.05. Saline-injected SHRs showed a tendency to respond more variably than saline-injected WKYs, F(1, 23) = 3.926, p < 0.10, but amphetamine-injected SHRs did not differ statistically from saline-injected WKYs, F(1, 23) = 0.614, NS. Figure 3 (bottom) shows the percent vary acquisition data from Experiment 2b~ Strain, F(1, 23) = 22.888, p < 0.001, drug, F(l, 23) = 12.7i2, p < 0.01, and session, F(8, 184) = 12.426, p < 0.0001, effects were significant, with SHRs responding more variably than WKYs, and amphetamine causing a decrease in variability in both strains. None of the interactions reached significance. As in Experiment 2a, the saline-injected SHRs responded more variably than saline-injected WKYs, F(1, 23) = 9.501, p < 0.01, and, once again, amphetamine-injected SHRs did not differ from saline-injectedWKYs, F(1, 23) = 0.774, NS. A 2 x 2 x 3 ANOVA (strain x drug x sessions) on percent vary at asymptote showed no significant differences for either Experiments 2a or 2b, with all groups reaching the same level of variability at asymptote. In brief, nondrugged SHRs consistently responded more variably than nondrugged WKYs in Experiments 1, 2a, and 2b. On the other hand, whereas amphetamine generally increased vari-
MOOK AND NE(!RIN(;t!!R
g0-
q\
80-
,,.,,
z
_o I-
70"
<~ >
6O
5O
4O 0
i
!
i
!
2
4
6
8
i
10
i
12
100
0 o°0". 9O
i '°
-
"0,
\'%o
K
WKY-AMPH
----o--- WKY-SAL
7O
6O
-
5O 40
!
|
!
|
|
|
2
4
6
8
10
12
SESSIONS
FIG. 3. Percentage of trials in which SHRs and WKYs injected with amphetamine or saline varied their responses across sequences. The top graph shows Experiment 2a, the bottom shows Experiment 2b. ability in Experiment 1, it caused little effect in Experiment 2a and decreased variability in 2b, this being seen in both SHRs and WKYs. Furthermore, whereas the amphetamine-administered SHRs responded more variably than the saline-administered WKYs in Experiment l, these two groups did not differ in Experiments 2a or 2b. This pattern of results will be discussed below. GENERAL DISCUSSION Previous research has shown that when the SHR strain of rats must learn to repeat a given response (14), or repeatedly select a subset of response sequences (16), they are less proficient than control WKY rats. The results of Experiment 2 support this claim. Given 16 possible response sequences, saline4njected SHRs took significantly longer than saline-injected WKYs to learn the four potentially reinforced sequences. Previous research has also shown that when SHRs and WKYs are rewarded for varying their responses, the SHRs perform at higher levels of proficiency than control WKYs (14,16). Experiment 1 found that when variable sequences of left and right responses were required for reinforcement, saline-injected SHRs varied their responses significantly more than saline-injected WKYs. Similar results were obtained in Experiments 2a and 2b. Thus, consistent with previous findings (14,16), WKYs were more readily reinforced for repeating whereas SHRs were more readily reinforced for varying. Strain of rats interacted with contingencies of reinforcement to determine whether SHRs showed a learning deficit or learning advantage. The main question asked in the present research was whether amphetamine differentially affected performances of SHRs vs.
SHR AND AMPHETAMINE
943
WKYs; whether there was a differential effect of the drug on ability to vary vs. repeat; or whether the effects of amphetamine depended upon an interaction between strain and reinforcement contingencies. In Experiment 2, amphetamine improved SHR's learning to repeat a subset of sequences but the drug had little effect on WKYs. Consequently, the amphetamine-injected SHRs learned to select the correct sequences as rapidly as the nondrugged WKYs. Thus, whereas nondrugged SHRs took longer to learn to repeat the four sequences than the nondrugged WKYs, amphetamine enabled the SHRs to approximate the control subjects' level of learning. These results may be analogous to the beneficial effects reported from administering methylphenidate and other stimulants to children diagnosed with ADHD (e.g., with respect to such repetition-demanding school tasks as spelling, arithmetic, and word memory) (8). The effects of amphetamine on rewarded variability are more difficult to interpret. Amphetamine increased variability in both strains in Experiment 1, had no effect in Experiment 2a, and decreased variability in both strains in Experiment 2b. Of most importance for a comparison with human ADHD, amphetamine-injected SHRs differed significantly from nondrugged controls in Experiment 1 but not in Experiments 2a or 2b. One possible explanation of these different effects of amphetamine on variability involves changes in sensitivity to the amphetamine across experiments. Future research should test for order-of-presentation and sensitization effects. In addition, d o s e - r e s p o n s e curves may differ for SHR vs. WKY, or for rewarded variations vs. repetitions, or both (4), and future research should systematically examine the effects of amphetamine dose on the present tasks. A third possible explanation involves the variability contingency itself. In Experiment 1, the lag 4 contingencies required that the current sequence differ from each of the previous four
sequences, whereas in Experiments 2a and 2b, the lag 1 contingencies required that the current sequence differ from only the single prior sequence. Other studies have shown that lag 4 contingencies lead to highly variable, stochastic-like performance (25), whereas lag 1 contingencies lead to alternation among two or three sequences, with such alternation probably involving memory for preceding sequences (16). Furthermore, in Experiment 1, all 16 sequences were potentially reinforced, whereas in Experiment 2, the subjects had to learn a subset, 4 of 16 sequences, again placing demands upon memory. Thus, the lag 4 contingencies in Experiment 1 may have resulted in subjects utilizing a variability strategy whereas the lag 1 contingencies in Experiment 2 may have led to utilization of a memory strategy. This third interpretation suggests that amphetamine administration caused SHRs to behave like control WKYs when performance was memory based, but that amphetamine did not narrow the difference between SHRs and controls when high levels of variation were demanded. If these and previous SHR results may be applied to the human sphere, then we predict that when highly variable behavior is adaptive, nondrugged ADHD individuals will outperform normal controls. We also predict that administration of amphetamine (Dexedrine), methylphenidate (Ritalin), or pemoline (Cylert) to the ADHD child will not decrease rewarded variability and therefore will not interfere with the ADHD child's advantage. These predictions should, of course, be tested.
ACKNOWLEDGEMENTS We thank Gene Olson for his care of the animals, advice, and assistance. This work was partly derived from an undergraduate senior thesis submitted by D.M.M. to Reed College and partly supported by National Science Foundation grant BNS-8707992.
REFERENCES 1. Barkley, R. A. Attention deficit hyperactivity disorder. New York: Guilford Press; 1990. 2. Berger, D. F.; Starzec, J. J. Contrasting lever-press avoidance behaviors of spontaneously hypertensive and normotensive rats (Rattus norvegicus). J. Comp. Psychol. 102(3):279-286; 1988. 3. Biederman, J.; Newcorn, J.; Sprich, S. Comorbidity of attention deficit hyperactivity disorder with conduct, depressive, anxiety, and other disorders. Am. J. Psychiatry 148(5):564-577; 1991. 4. Blackman, D. E.; Pelion, R. The contributions of B. F, Skinner to the interdisciplinary science of behavioural pharmacology. Br. J. Psychol. 84:1-25; 1993. 5. Cantwell, D.; Baker, L. Association between attention deficit-hyperactivity disorder and learning disorders. J. Learn. Disab. 24(2):88-95; 1991. 6. Cierpiai, M. A.; Shasby, D. E.; Murphy, C. A.; et al. Open-field behavior of spontaneously hypertensive and Wistar-Kyoto normotensive rats: Effects of reciprocal cross-fostering. Behav. Neural Biol. 51:203-210; 1989. 7. Danysz, W.; Plaznik, A.; Pucilowski, O.; Plewako, M.; Obersztyn, M.; Kostowski, W. Behavioral studies in spontaneously hypertensive rats. Behav. Neural Biol. 39:22-29; 1983. 8. Douglas, V. I.; Barr, R. G.; O'Neil, M. E.; Britton, B. G. Short term effects of methylphenidate on the cognitive, learning and academic performance of children with attention deficit disorder in the laboratory and the classroom. J. Child Psychol. Psychiatry 27(2): 191211; 1985. 9. Gentsch, C.; Lichtsteiner, M.; Feer, H. Open field and elevated plus-maze: A behavioral comparison between spontaneously hypertensive (SHR) and Wistar-Kyoto (WKY) rats and the effect of chlordiazepoxide. Behav. Brain Res. 25:101-107; 1987.
10. Huang, K. H.; Shyu, B. C. Differential stress effects on responses to noxious stimuli as measured by tail-flick latency and squeak threshold in rats. Acta Physiol. Scand. 129:401-406; 1987. 11. Knardahl, S. Behavioral and psychophysiologicai characteristics of the spontaneously hypertensive rat. Acta Med. Scand. Suppl. 714:139-145; 1986. 12. Knardahl, S.; Sagvolden, T. Open-field behavior of spontaneously hypertensive rats. Behav. Neural Biol. 27:187-200; 1979. 13. Kupietz, S.; Winsberg, B.; Richardson, E.; Maitinsky, S.; Mendeil, N. Effects of methylphenidate dosage in hyperactive readingdisabled children: I. Behavior and cognitive performance effects. J. Am. Acad. Child Adolesc. Psychiatry 27(1):70-77; 1988. 14. Low, W.; Whitehorn, D.; Hendley, E. Genetically related rats with differences in hippocampal uptake of norepinephrine and maze performance. Brain Res. Bull. 12(6):703-709; 1984. 15. Maixner, W.; Touw, K. B.; Brody, M. J.; Gebhart, G. F.; Long, J. P. Factors influencing the altered pain perception in the spontaneously hypertensive rat. Brain Res. 237:137-145; 1982. 16. Mook, D.; Jeffrey, J.; Neuringer, A. Behavioral variation and selection by spontaneously hypertensive (SHR) and normotensive (WKY) rats. Behav. Neural Biol. 59:126-135; 1993. 17. Morgan, L.; Neuringer, A. Behavioral variability as a function of response topography and reinforcement contingency. Anim. Learn. Behav. 18:257-263; 1990. 18. Moser, M. B.; Moser, E. I.; Wultz, B.; Sagvoiden, T. Component analyses differentiate between exploratory behavior of spontaneously hypertensive rats and Wistar-Kyoto rats in a two-compartment free-exploration open field. Scand. J. Psychol. 29:200-206; 1988.
944
19. Myers, M. M.; Musty, R. E.; Hendley, E. D. Attenuation of hyperactivity in the spontaneously hypertensive rats by amphetamine. Behav. Neural Biol. 34:42-54; 1982. 20. Neuringer, A. Operant variability and repetition as functions of interresponse time. J. Exp. Psychol. [Anim. Behav.] 17:3-- 12; 199 I. 21. Neuringer, A. Reinforced variation and selection. Anita. Learn. Behav. 21:83-91; 1993. 22. Neuringer, A.; Huntley, R. W. Reinforced variability in rats: Effects of gender, age and contingency. Physiol. Behav. 51:145149; 1992. 23. Nomura, M. Effect of nimodipine on brightness discrimination learning test in Wistar-Kyoto and spontaneously hypertensive rats. Drug Res. 38(I1)(9): 1282-1286; 1988. 24. Okamoto, K. Spontaneous hypertension in rats. Int. Rev. Exp. Pathol. 7:222-268; 1969. 25. Page, S.; Neuringer, A. Variability is an operant. J. Exp. Psychol. [Anim. Behav.] 11:429-452; 1985. 26. Rapport, M.; Quinn, S.; DuPaul, G.; Quinn, E.; Kelley, K. Attention deficit disorder with hyperactivity and methylphenidate: The effects of dose and mastery level on children's learning performance. J. Abnorm. Child Psychol. 17:669-689; 1989.
MOOK AND N[';URINGER
27. Rasbury, W. Attention deficit disorder: An overview. Henry l-:ord Hosp. Med. J. 36(4):212-216; 1988. 28. Rodgers, L. J.; Sink, H. S.; Hambly, J. W. Exploration, lear and maze learning in spontaneously hypertensive and nonnotensive rats. Behav. Neural Biol. 49:222-223; 1988. 29. Sagvolden, T.; Metzger, M.; Schiorbeck, H.; Rugland, A.; Spinnanget, 1.; Sagvolden, G. The spontaneously hypertensive rat (SHR) as an animal model of childhood hyperactivity (ADHD): Changed reactivity to reinforcers and to psychomotor stimulants. Behav. Neural Biol. 58:103-112; 1992. 30. Sagvolden, T.; Metzger, M.; Sagvolden, G. Frequent reward eliminates differences in activity between hyperkinetic rats and controls. Behav. Neural Biol. 59:225-229; 1993. 31. Van den Buuse, M.; de Jong, W. Differential effects of dopaminergic drugs on open field behavior of spontaneously hypertensive rats and normotensive Wistar-Kyoto rats. J. Pharmacol. Exp. Ther~ 248(3): 1189-1196; 1989. 32. Wultz, B.; Sagvolden, T.; Moser, E. 1.; Moser, M. B. The spontaneously hypertensive rat as an animal model of attention-deficit hyperactivity disorder: Effects of methylphenidate on exploratory behavior. Behav. Neural Biol. 53:88-102; 1990.
Pergamon 0031-9384(94)00160-X
Different Effects of Amphetamine on Reinforced Variations Versus Repetitions in Spontaneously Hypertensive Rats (SHR) DEBORAH
M. MOOK
AND
ALLEN
NEURINGER I
Department o f Psychology, Reed College, Portland, OR 97202 R e c e i v e d 23 A u g u s t 1993
MOOK, D. M. AND A. NEURINGER. Different effects of amphetamine on reinforced variations versus repetitions in spontaneously hypertensive rats (SHR). PHYSIOL BEHAV $6(5) 939-944, 1994.--The spontaneously hypertensive rat (SHR) may serve as an animal model of human attention deficit hyperactivity disorder (ADHD). We compared performances of SHRs and Wystar-Kyoto normotensive controls rats (WKY) in two experiments. When rewarded for varying sequences of responses across two manipulanda, the SHRs were more likely to vary than the WKYs. On the other hand, when rewarded for repetitions of a small number of sequences, the WKYs were more likely to learn to repeat. Both of these results confirm previous findings. Injecting 0.75 mg/kg d-amphetamine facilitated learning by SHRs to repeat the required sequences, with amphetamine-injected SHRs learning as rapidly as saline-injected, control WKYs. On the other hand, amphetamine tended to increase variability in both strains when high levels of variations were required for reward, and to decrease it in both strains when low levels of variability were required. Thus, amphetamine may have different effects on reinforced repetitions vs. reinforced variations. SHR rats ADHD Behavioral variability Amphetamine Reinforcement Operant conditioning Selection
THE spontaneously hypertensive rat (SHR), originally studied because of its high blood pressure (24), may also serve as an animal model for another human disorder, attention deficit hyperactivity disorder, or A D H D (29,32). There are two main sources of support for this hypothesis. First, SHRs are more active than their normotensive W i s t a r - K y o t o (WKY) progenitors, typically used as controls (6,9,10,16,31,32). Children with A D H D have also been shown to be more active than controls (26,27). Second, SHRs take longer to learn some tasks than WKYs and learn to lower levels of proficiency (14,23). A D H D children may also show such learning deficits (1,3,5,26). Thus, high levels of activity and learning disabilities have been reported both for children diagnosed with A D H D and SHR rats ( 1,3,5,6, 8,9,12,13,18,26,27,30- 32). SHRs do not always show learning deficits, however. For example, they avoid or escape aversive stimuli more readily than WKYs (7,11). This finding may partly be accounted for by the fact that aversive stimuli elicit different responses in the two s t r a i n s - - S H R s tend to move and W K Y s to freeze (2). Sensitivity to painful stimulation may also differ in the two strains [(15), see, however, (l 0)]. Even under some positively reinforcing conditions, SHRs learn more rapidly than W K Y s (e.g., in the H e b b -
Hyperactivity
Response sequences
Memory
William's maze) (28). Thus, a general learning deficit does not appear to characterize SHR performance. Two studies may help to explain why SHRs sometimes perform less successfully than WKYs, while at other times they are more successful. The original clue was provided by Low, Whitehorn, and Hendley (14). When reinforcement depended on rats learning to go to a single arm of a T-maze, WKYs were more proficient than SHRs; but when reinforcement depended on rats varying their arm entries, SHRs excelled. Mook, Jeffrey, and Neuringer (16) extended the Low et al. (14) findings with two different procedures. (a) In a 12-arm radial maze task, where variations of arm entries are required, SHRs varied their arm choices more, making fewer repetition errors, than WKYs. On the other hand, when the requirement for response variability was combined with a requirement to repeat selected arms in the radial maze, WKYs more rapidly learned to repeat the required arms. (b) Similar results were obtained in an operant chamber: when reinforcement depended on variable sequences of left and fight responses across two levers, the SHRs responded more variably than WKYs, but when sequences had to be repeated, the WKYs better learned to repeat. Thus, in both the Low et al. and Mook et al. studies, strain of rats interacted with reinforcement contin-
To whom requests for reprints should be addressed. 939
940
MOOK AND NEURIN(;ER
gencies in determining whether SHRs showed a learning deficit or advantage. When required to vary, SHRs had an advantage over WKYs; when required to repeat, WKYs had the advantage. ADHD in humans is commonly treated with stimulant medication (e.g., Dexedrine, Ritalin, or Cylert), which causes diagnosed subjects to behave in a manner approximating that of normal controls (1,8,12,26,27). Several studies have explored the effects of stimulant medication on SHR's locomotor responses and response rates (19,29-32). The present study extends that research by asking whether amphetamine has different effects when response variations vs. repetitions are required. The general question is whether the effects of amphetamine on SHRs vs. WKYs depend upon reinforcement contingencies. GENERAL METHOD
Subjects Fourteen male SHR and 14 male WKY, obtained from Harlan Sprague-Dawley, Inc. (Indianapolis, IN), were housed in pairs in 40.5 x 24 × 19 cm cages under a 12-h light/dark cycle, and maintained on a 21-h food-deprivation schedule, with water available freely except during the experimental hour. The same subjects, experimentally naive at the beginning, were used in all experiments.
Apparatus Nine identical modified Gerbrands operant chambers were 27 cm deep, 28 cm long, and 30 cm high. The ceiling, back wall, and front wall (which opened on hinges) were constructed of clear Plexiglas, and the two side walls were constructed of aluminum. The floor consisted of wire bars 1 cm apart. One side wall contained three pigeon response keys that were 3 cm in diameter, 17.5 cm from the ceiling, and 9.5 cm from the floor. The keys were 9 cm apart from center to center, and could be illuminated by white 24-V bulbs. An audio speaker was located behind the wall containing the keys. Only two keys were act i v e - t h e middle key and the right key. On the opposite wall were two Gerbrands response levers that were not used in this experiment. Located between the two levers was a pellet tray, 3 cm from the floor, into which 45-mg Noyes pellets could be released. Each chamber was housed in a sound- and light-attenuating outer cubicle made of wood and connected to an Apple Macintosh computer through a Metaresearch Benchtop interface. These cubicles were equipped with one-way mirrors through which subjects could be observed during sessions.
trial consisted of lour pushes on left (L) or right (R) keys. If the current sequence differed from each of the previous R~ur sequences (a lag 4 variability contingency), then a lk)od pellet was presented. Thus, for example, if the prior sequences had been LRRR, RRRR, RLLR, and RRRR, then an LLRL sequence would end with reinforcement, because it differed flom each of the previous four, whereas an RLLR sequence would end with a brief time-out, because it repeated one of the previous four sequences. METHOD
Procedure Subjects were first autoshaped to press each of the two keys for food reinforcement, and then placed on a fixed ratio (FR) schedule, the value of the ratio increasing to FR4 over the course of four sessions. Four responses summed across the keys were required for reinforcement. The variability schedule was then introduced, in which four presses on left and right keys constituted a trial, and the current sequence was required to differ from each of the sequences in the preceding four trials for the animal to be reinforced, a lag 4, variability contingency (17,25). As in previous studies, the four-trial (or lag) window moved at the end of each trial, so that the current sequence always had to differ from the just-prior four sequences for reinforcement to be delivered. When the variability contingency had been met, a food pellet was presented, preceded by a 0.5-s end-of-trial cue, namely darkening of the chamber and a 200 Hz tone alternating with a 1000 Hz tone, creating a warbling sound. If the sequence in the current trial repeated any one or more of the previous four sequences, then a 3-s time-out was imposed, consisting of the same dark chamber but with no tone and no reinforcer. After time-out or reinforcement, the two keys were reilluminated and a new trial began. Sessions terminated after 100 trials or 0.5 h, whichever occurred first. Ten sessions were provided, 5 days per week.
Data Analysis Percent vary, the main dependent variable, was calculated by dividing the number of sequences in a session that differed from each of the previous four sequences (i.e., number of reinforced sequences) by the total number of sequences (or total trials)in the session, the result multiplied by 100 to obtain a percentage. RESULTS AND DISCUSSION Figure 1 shows group average percent vary across 10 sessions. A 2 x 2 × 10 (strain X drug x sessions) ANOVA showed that
Drugs and Dosage d-Amphetamine sulfate was dissolved in 0.9% saline and injected IP in a dose of 0.75 mg/kg by body weight. Fifteen minutes prior to each session, all animals were weighed and administered with either the drug solution or with an equal volume of 0.9% saline according to group assignment. Assignment to amphetamine vs. saline conditions was random, with seven SHRs and seven WKYs receiving saline and seven SHRs and seven WKYs receiving amphetamine. Thus, the four experimental groups were: SHR-amphetamine (SHR-AMPH), SHR-saline (SHRSAL), WKY-amphetamine (WKY-AMPH), and WKY-saline WKY-SAL). EXPERIMENT 1: OPERANT VARIABILITY Variable sequences of left and right responses were reinforced in an operant chamber, a task similar to that previously used with rats and pigeons [e.g., (16,17,20,22,25)]. In the present case, a
8070-
tr ,~
50
4O 30
WKY..SAL
---"0--
!
!
!
2
4
6
-
|
8
. "'"
I
10
SESSIONS
FIG. 1. Percentage of reinforcements for SHRs and WKYs given amphetamine and saline under the variability contingency during each of 10 sessions.
SHR AND AMPHETAMINE
there was a significant strain effect, F(1, 24) = 8.569, p < 0.01, with SHRs responding more variably than WKYs; a weak drug effect, F(1, 24) = 3.622, p < 0.10, with subjects given amphetamine responding more variably than those given saline; and a significant sessions effect, F(9, 216) = 3.954, p < 0.001, with variability increasing across sessions. None of the interactions reached significance. Two additional planned comparisons were performed to test (i) whether nondrugged SHRs responded more variably than nondrugged WKYs, a finding reported in previous studies, and (ii) whether amphetamine administered to SHRs would cause them to behave similarly to control, nondrugged WKYs, a comparison relevant to treatment of human ADHD with stimulant medication. We found that (i) saline-injected SHRs were indeed significantly more variable than similarly injected WKYs, F(1, 24) = 8.273, p < 0.01, results consistent with previous findings (14,16). However, (ii) amphetamine administered to SHRs did not cause them to perform like saline-injected WKYs, with the SHRAMPH group responding significantly more variably than the WKY-SAL group, F(1, 24) = 11.667, p < 0.01. This comparison therefore did not support the hypothesis that amphetamine causes SHRs, a proposed animal model of human ADHD, to behave like normal animals. The next experiment further tested these questions in a situation where both variations and repetitions were required. EXPERIMENT 2: OPERANT REPETITION AND VARIATION Experiment 2 employed the Mook et al. (16) procedure. Rats were required to select repeatedly 4 of 16 possible sequences, but could not repeat the same sequence emitted in the just-previous trial (i.e. a lag 1 variability contingency was imposed within the reinforcable subset). The main goal was to test the Experiment 1 findings that amphetamine affected SHRs and WKYs similarly and that amphetamine-injected SHRs differed significantly from saline-injected control subjects.
941
lection was calculated by dividing the number of within-subset sequences by the total number of sequences, multiplied by 100. The numerator included both reinforced and nonreinforced (or repeated) entries into the designated subset. Percent vary was calculated by dividing the number of reinforced sequences by total number of emissions of the subset, multiplied by 100. Thus, a high value for percent selection indicated that the rat learned to emit the four potentially reinforced sequences and to avoid emitting the other sequences. A high value for percent vary indicated that many within-subset sequences were preceded by a different sequence, either a different within-subset sequence or any nonsubset sequence. RESULTS AND DISCUSSION
Selection Figure 2 (top) shows group average percent selection across the 12 sessions of Experiment 2a. The figure indicates that performances approached asymptote at about the ninth session. The data were therefore divided into two parts for ease of analysis: acquisition (nine sessions of learning) and asymptote (three final sessions). A 2 × 2 × 9 ANOVA (strain × drug × sessions) on percent selection (acquisition) showed a significant effect of strain, F(I, 23) = 7.553, p < 0.05, with WKYs selecting the correct subset more frequently than SHRs; a significant effect of drug, F(1, 23) = 5.008, p < 0.05, with amphetamine-injected rats selecting the correct subset more frequently than saline-injected; and a significant effect of sessions, F(8, 184) = 28.639, p < 0.0001, with performances improving across sessions. None of the interactions approached significance. The first of two additional planned com80 70 ¸
METHOD
Procedure As in Experiment 1, four responses to the two keys constituted a trial. However, in the present experiment, only four sequences could be reinforced with the other 12 sequences never reinforced. The potentially reinforced sequences in Experiment 2a, defined as the subset, were RRRR, RRRL, RRLR, and RRLL. A sequence in this subset was reinforced only if it differed from the previously emitted sequence (i.e., a lag 1 variability contingency). To optimize reinforcement, subjects therefore had to learn: (a) to emit only the four sequences in the subset, and (b) not to emit the same sequence twice in a row. Subjects were run for 12 sessions, generally five sessions per week, with each session terminating after 100 trials or 0.5 h, whichever occurred first. Experiment 2b repeated the procedure with one difference. The subset was changed to LLLL, LLLR, LLRL, and LLRR (i.e., the mirror-image sequences of those used in Experiment 2a). As will be seen, the results of Experiment 2a were of some interest, and Experiment 2b attempted to replicate 2a with sequences that were most comparable in difficulty but for which there would be minimal direct transfer [see (21)]. Experiment 2b followed immediately after 2a.
Data Analysis The two dependent measures were percent selection of the required subset and percent vary within the subset. Percent se-
5O LU
¢/)
40
~
3o 20
10 0
80 70 60 50 40
3o
,,O"
,'~ • ;/"
"~'O'''O'- .O"-
-
-
~-
=u,~_AuI SHR-AMPH
20 10
I
i
l
I
2
4
6
8
I0
1
I
12
SESSIONS FIG. 2. Percentage of trials in which SHRs and WKYs injected with amphetamine or saline emitted any of the four poteniaUy reinforced response sequences. The top graph shows Experiment 2a, the bottom shows Experiment 2b.
942
parisons showed that saline-injected WKYs selected the correct subset more frequently than saline-injected SHRs, F(1, 23) 5.656, p < 0.05, again replicating the previous findings. The second planned comparison showed that SHRs injected with amphetamine were statistically indistinguishable from control WKYs administered saline, F(l, 23) = 0.136, NS. The improvement due to amphetamine approached significance for SHRs (p < 0.06), but was not statistically significant for WKYs. Thus, unlike Experiment l, but consistent with human ADHD findings, amphetamine caused SHRs to approximate the performance of nondrugged control subjects. Similar results were obtained in Experiment 2b, as shown in Fig. 2 (bottom). There were significant strain, F(1,23) = 18.993, p < 0.001, drug, F(l, 23) -- 5.165,p < 0.05, and sessions effects, F(8, 184) = 54.464, p < 0.0001, and, in this case, a significant drug by strain interaction, F(1, 23) = 5.496, p < 0.05. As in Experiment 2a, the saline-injected WKYs learned more rapidly than the saline-injected SHRs, F(l, 23) = 23.397, p < 0.0001, and injecting the SHRs with amphetamine caused them to learn as rapidly as the control WKYs, F(1, 23) = 2.265, NS. Again as in Experiment 2a, amphetamine facilitated learning to repeat by the SHRs, F(1, 23) = I 1.02, p < 0.01, but not the WKYs, F(l, 23) = 0.002, NS. Thus, Experiment 2b essentially replicated the results of Experiment 2a with respect to repetitions of subsets of responses. Two 2 × 2 × 3 ANOVAs (strain × drug × sessions) on percent selection at asymptote (i.e., during the last three sessions of the experiments) yielded no significant differences. In each of the two experiments, the groups attained levels of performance that did not differ at asymptote. Variation
Figure 3 (top) shows percent vary within the subset in Experiment 2a. The data were again divided into acquisition and asymptote phases. A 2 X 2 x 9 ANOVA (strain × drug x sessions) of the percent vary (acquisition) data showed a weak strain effect, F(1, 23) = 3.017, p < 0.10, with SHRs behaving more variably than WKYs, no significant effect of drug, and a significant sessions effect, F(8, 184) = 12.449,p < 0.0001. The strain by sessions interaction was significant, F(8, 184) = 2.758, p < 0.01, with SHRs behaving more variably than WKYs during the initial sessions. The strain by drug by sessions interaction also attained significance, F(8, 184) = 2.024, p < 0.05. Saline-injected SHRs showed a tendency to respond more variably than saline-injected WKYs, F(1, 23) = 3.926, p < 0.10, but amphetamine-injected SHRs did not differ statistically from saline-injected WKYs, F(1, 23) = 0.614, NS. Figure 3 (bottom) shows the percent vary acquisition data from Experiment 2b~ Strain, F(1, 23) = 22.888, p < 0.001, drug, F(l, 23) = 12.7i2, p < 0.01, and session, F(8, 184) = 12.426, p < 0.0001, effects were significant, with SHRs responding more variably than WKYs, and amphetamine causing a decrease in variability in both strains. None of the interactions reached significance. As in Experiment 2a, the saline-injected SHRs responded more variably than saline-injected WKYs, F(1, 23) = 9.501, p < 0.01, and, once again, amphetamine-injected SHRs did not differ from saline-injectedWKYs, F(1, 23) = 0.774, NS. A 2 x 2 x 3 ANOVA (strain x drug x sessions) on percent vary at asymptote showed no significant differences for either Experiments 2a or 2b, with all groups reaching the same level of variability at asymptote. In brief, nondrugged SHRs consistently responded more variably than nondrugged WKYs in Experiments 1, 2a, and 2b. On the other hand, whereas amphetamine generally increased vari-
MOOK AND NE(!RIN(;t!!R
g0-
q\
80-
,,.,,
z
_o I-
70"
<~ >
6O
5O
4O 0
i
!
i
!
2
4
6
8
i
10
i
12
100
0 o°0". 9O
i '°
-
"0,
\'%o
K
WKY-AMPH
----o--- WKY-SAL
7O
6O
-
5O 40
!
|
!
|
|
|
2
4
6
8
10
12
SESSIONS
FIG. 3. Percentage of trials in which SHRs and WKYs injected with amphetamine or saline varied their responses across sequences. The top graph shows Experiment 2a, the bottom shows Experiment 2b. ability in Experiment 1, it caused little effect in Experiment 2a and decreased variability in 2b, this being seen in both SHRs and WKYs. Furthermore, whereas the amphetamine-administered SHRs responded more variably than the saline-administered WKYs in Experiment l, these two groups did not differ in Experiments 2a or 2b. This pattern of results will be discussed below. GENERAL DISCUSSION Previous research has shown that when the SHR strain of rats must learn to repeat a given response (14), or repeatedly select a subset of response sequences (16), they are less proficient than control WKY rats. The results of Experiment 2 support this claim. Given 16 possible response sequences, saline4njected SHRs took significantly longer than saline-injected WKYs to learn the four potentially reinforced sequences. Previous research has also shown that when SHRs and WKYs are rewarded for varying their responses, the SHRs perform at higher levels of proficiency than control WKYs (14,16). Experiment 1 found that when variable sequences of left and right responses were required for reinforcement, saline-injected SHRs varied their responses significantly more than saline-injected WKYs. Similar results were obtained in Experiments 2a and 2b. Thus, consistent with previous findings (14,16), WKYs were more readily reinforced for repeating whereas SHRs were more readily reinforced for varying. Strain of rats interacted with contingencies of reinforcement to determine whether SHRs showed a learning deficit or learning advantage. The main question asked in the present research was whether amphetamine differentially affected performances of SHRs vs.
SHR AND AMPHETAMINE
943
WKYs; whether there was a differential effect of the drug on ability to vary vs. repeat; or whether the effects of amphetamine depended upon an interaction between strain and reinforcement contingencies. In Experiment 2, amphetamine improved SHR's learning to repeat a subset of sequences but the drug had little effect on WKYs. Consequently, the amphetamine-injected SHRs learned to select the correct sequences as rapidly as the nondrugged WKYs. Thus, whereas nondrugged SHRs took longer to learn to repeat the four sequences than the nondrugged WKYs, amphetamine enabled the SHRs to approximate the control subjects' level of learning. These results may be analogous to the beneficial effects reported from administering methylphenidate and other stimulants to children diagnosed with ADHD (e.g., with respect to such repetition-demanding school tasks as spelling, arithmetic, and word memory) (8). The effects of amphetamine on rewarded variability are more difficult to interpret. Amphetamine increased variability in both strains in Experiment 1, had no effect in Experiment 2a, and decreased variability in both strains in Experiment 2b. Of most importance for a comparison with human ADHD, amphetamine-injected SHRs differed significantly from nondrugged controls in Experiment 1 but not in Experiments 2a or 2b. One possible explanation of these different effects of amphetamine on variability involves changes in sensitivity to the amphetamine across experiments. Future research should test for order-of-presentation and sensitization effects. In addition, d o s e - r e s p o n s e curves may differ for SHR vs. WKY, or for rewarded variations vs. repetitions, or both (4), and future research should systematically examine the effects of amphetamine dose on the present tasks. A third possible explanation involves the variability contingency itself. In Experiment 1, the lag 4 contingencies required that the current sequence differ from each of the previous four
sequences, whereas in Experiments 2a and 2b, the lag 1 contingencies required that the current sequence differ from only the single prior sequence. Other studies have shown that lag 4 contingencies lead to highly variable, stochastic-like performance (25), whereas lag 1 contingencies lead to alternation among two or three sequences, with such alternation probably involving memory for preceding sequences (16). Furthermore, in Experiment 1, all 16 sequences were potentially reinforced, whereas in Experiment 2, the subjects had to learn a subset, 4 of 16 sequences, again placing demands upon memory. Thus, the lag 4 contingencies in Experiment 1 may have resulted in subjects utilizing a variability strategy whereas the lag 1 contingencies in Experiment 2 may have led to utilization of a memory strategy. This third interpretation suggests that amphetamine administration caused SHRs to behave like control WKYs when performance was memory based, but that amphetamine did not narrow the difference between SHRs and controls when high levels of variation were demanded. If these and previous SHR results may be applied to the human sphere, then we predict that when highly variable behavior is adaptive, nondrugged ADHD individuals will outperform normal controls. We also predict that administration of amphetamine (Dexedrine), methylphenidate (Ritalin), or pemoline (Cylert) to the ADHD child will not decrease rewarded variability and therefore will not interfere with the ADHD child's advantage. These predictions should, of course, be tested.
ACKNOWLEDGEMENTS We thank Gene Olson for his care of the animals, advice, and assistance. This work was partly derived from an undergraduate senior thesis submitted by D.M.M. to Reed College and partly supported by National Science Foundation grant BNS-8707992.
REFERENCES 1. Barkley, R. A. Attention deficit hyperactivity disorder. New York: Guilford Press; 1990. 2. Berger, D. F.; Starzec, J. J. Contrasting lever-press avoidance behaviors of spontaneously hypertensive and normotensive rats (Rattus norvegicus). J. Comp. Psychol. 102(3):279-286; 1988. 3. Biederman, J.; Newcorn, J.; Sprich, S. Comorbidity of attention deficit hyperactivity disorder with conduct, depressive, anxiety, and other disorders. Am. J. Psychiatry 148(5):564-577; 1991. 4. Blackman, D. E.; Pelion, R. The contributions of B. F, Skinner to the interdisciplinary science of behavioural pharmacology. Br. J. Psychol. 84:1-25; 1993. 5. Cantwell, D.; Baker, L. Association between attention deficit-hyperactivity disorder and learning disorders. J. Learn. Disab. 24(2):88-95; 1991. 6. Cierpiai, M. A.; Shasby, D. E.; Murphy, C. A.; et al. Open-field behavior of spontaneously hypertensive and Wistar-Kyoto normotensive rats: Effects of reciprocal cross-fostering. Behav. Neural Biol. 51:203-210; 1989. 7. Danysz, W.; Plaznik, A.; Pucilowski, O.; Plewako, M.; Obersztyn, M.; Kostowski, W. Behavioral studies in spontaneously hypertensive rats. Behav. Neural Biol. 39:22-29; 1983. 8. Douglas, V. I.; Barr, R. G.; O'Neil, M. E.; Britton, B. G. Short term effects of methylphenidate on the cognitive, learning and academic performance of children with attention deficit disorder in the laboratory and the classroom. J. Child Psychol. Psychiatry 27(2): 191211; 1985. 9. Gentsch, C.; Lichtsteiner, M.; Feer, H. Open field and elevated plus-maze: A behavioral comparison between spontaneously hypertensive (SHR) and Wistar-Kyoto (WKY) rats and the effect of chlordiazepoxide. Behav. Brain Res. 25:101-107; 1987.
10. Huang, K. H.; Shyu, B. C. Differential stress effects on responses to noxious stimuli as measured by tail-flick latency and squeak threshold in rats. Acta Physiol. Scand. 129:401-406; 1987. 11. Knardahl, S. Behavioral and psychophysiologicai characteristics of the spontaneously hypertensive rat. Acta Med. Scand. Suppl. 714:139-145; 1986. 12. Knardahl, S.; Sagvolden, T. Open-field behavior of spontaneously hypertensive rats. Behav. Neural Biol. 27:187-200; 1979. 13. Kupietz, S.; Winsberg, B.; Richardson, E.; Maitinsky, S.; Mendeil, N. Effects of methylphenidate dosage in hyperactive readingdisabled children: I. Behavior and cognitive performance effects. J. Am. Acad. Child Adolesc. Psychiatry 27(1):70-77; 1988. 14. Low, W.; Whitehorn, D.; Hendley, E. Genetically related rats with differences in hippocampal uptake of norepinephrine and maze performance. Brain Res. Bull. 12(6):703-709; 1984. 15. Maixner, W.; Touw, K. B.; Brody, M. J.; Gebhart, G. F.; Long, J. P. Factors influencing the altered pain perception in the spontaneously hypertensive rat. Brain Res. 237:137-145; 1982. 16. Mook, D.; Jeffrey, J.; Neuringer, A. Behavioral variation and selection by spontaneously hypertensive (SHR) and normotensive (WKY) rats. Behav. Neural Biol. 59:126-135; 1993. 17. Morgan, L.; Neuringer, A. Behavioral variability as a function of response topography and reinforcement contingency. Anim. Learn. Behav. 18:257-263; 1990. 18. Moser, M. B.; Moser, E. I.; Wultz, B.; Sagvoiden, T. Component analyses differentiate between exploratory behavior of spontaneously hypertensive rats and Wistar-Kyoto rats in a two-compartment free-exploration open field. Scand. J. Psychol. 29:200-206; 1988.
944
19. Myers, M. M.; Musty, R. E.; Hendley, E. D. Attenuation of hyperactivity in the spontaneously hypertensive rats by amphetamine. Behav. Neural Biol. 34:42-54; 1982. 20. Neuringer, A. Operant variability and repetition as functions of interresponse time. J. Exp. Psychol. [Anim. Behav.] 17:3-- 12; 199 I. 21. Neuringer, A. Reinforced variation and selection. Anita. Learn. Behav. 21:83-91; 1993. 22. Neuringer, A.; Huntley, R. W. Reinforced variability in rats: Effects of gender, age and contingency. Physiol. Behav. 51:145149; 1992. 23. Nomura, M. Effect of nimodipine on brightness discrimination learning test in Wistar-Kyoto and spontaneously hypertensive rats. Drug Res. 38(I1)(9): 1282-1286; 1988. 24. Okamoto, K. Spontaneous hypertension in rats. Int. Rev. Exp. Pathol. 7:222-268; 1969. 25. Page, S.; Neuringer, A. Variability is an operant. J. Exp. Psychol. [Anim. Behav.] 11:429-452; 1985. 26. Rapport, M.; Quinn, S.; DuPaul, G.; Quinn, E.; Kelley, K. Attention deficit disorder with hyperactivity and methylphenidate: The effects of dose and mastery level on children's learning performance. J. Abnorm. Child Psychol. 17:669-689; 1989.
MOOK AND N[';URINGER
27. Rasbury, W. Attention deficit disorder: An overview. Henry l-:ord Hosp. Med. J. 36(4):212-216; 1988. 28. Rodgers, L. J.; Sink, H. S.; Hambly, J. W. Exploration, lear and maze learning in spontaneously hypertensive and nonnotensive rats. Behav. Neural Biol. 49:222-223; 1988. 29. Sagvolden, T.; Metzger, M.; Schiorbeck, H.; Rugland, A.; Spinnanget, 1.; Sagvolden, G. The spontaneously hypertensive rat (SHR) as an animal model of childhood hyperactivity (ADHD): Changed reactivity to reinforcers and to psychomotor stimulants. Behav. Neural Biol. 58:103-112; 1992. 30. Sagvolden, T.; Metzger, M.; Sagvolden, G. Frequent reward eliminates differences in activity between hyperkinetic rats and controls. Behav. Neural Biol. 59:225-229; 1993. 31. Van den Buuse, M.; de Jong, W. Differential effects of dopaminergic drugs on open field behavior of spontaneously hypertensive rats and normotensive Wistar-Kyoto rats. J. Pharmacol. Exp. Ther~ 248(3): 1189-1196; 1989. 32. Wultz, B.; Sagvolden, T.; Moser, E. 1.; Moser, M. B. The spontaneously hypertensive rat as an animal model of attention-deficit hyperactivity disorder: Effects of methylphenidate on exploratory behavior. Behav. Neural Biol. 53:88-102; 1990.