d-amphetamine S EFFECTS ON BEHAVIOR PUNISHED BY TIME-OUT FROM POSITIVE REINFORCEMENT Emily E. Guido

Transcription

1 d-amphetamine S EFFECTS ON BEHAVIOR PUNISHED BY TIME-OUT FROM POSITIVE REINFORCEMENT Emily E. Guido A Thesis Submitted to the University of North Carolina Wilmington in Partial Fulfillment of the Requirements for the Degree of Master of Arts Department of Psychology University of North Carolina Wilmington 2010 Approved by Advisory Committee Raymond C. Pitts Anne Hungerford Wendy Donlin Washington Christine E. Hughes Chair Accepted by Dean, Graduate School

2 TABLE OF CONTENTS ABSTRACT... iv ACKNOWLEDGMENTS...v DEDICATION... vi LIST OF TABLES... vii LIST OF FIGURES... viii INTRODUCTION...1 d-amphetamine...1 Time-Out and Stimulant Prevalence...1 Stimulants and Punished Behavior...3 Shock Punishment and Stimulants...4 Time-Out and Stimulants...7 d-amphetamine and Schedule-Induced Behavior...9 d-amphetamine and Schedule-Controlled Behavior Punished by Time-Out...15 Current Experiment...16 METHOD...17 Subjects...17 Apparatus...18 Behavioral Procedure...19 Pharmacological Procedure...20 Data Analysis...21 RESULTS...22 DISCUSSION...33 ii

3 Time-Out and Punishment.33 d-amphetamine s Effects on Punished Schedule-Controlled Behavior 36 Rate-dependency, Timing, and d-amphetamine...39 Use of Time-Out and d-amphetamine with Children...43 REFERENCES...45 iii

4 ABSTRACT Amphetamine s effects on operant behavior punished by time-out have not been examined. Examining amphetamine s effects on operant behavior punished by time-out is crucial to understanding how time-out affects children who are taking stimulant medications. In the current study, a multiple random-interval (RI) 1-min RI 6-min schedule of food presentation was arranged for pigeons key pecking. Once behavior was stable under this schedule, a 20-s time-out was added to the RI 1-min component. The timeout was presented according to a random ratio 3 (RR3) for 2 pigeons and an RR2 for 2 pigeons. Time-out decreased rates in the RI 1-min component for all 3 pigeons. For 1 pigeon, the RI 6-min component rates increased. d-amphetamine had different effects across the 3 pigeons. There were rate-dependent effects for 1 pigeon with increases in the punished component and decreases in the unpunished component, a general decrease for another in both components, and an increase at low doses followed by a general decrease in response rates in both components for the other. iv

5 ACKNOWLEDGEMENTS I would like to thank my mentor, Dr. Christine Hughes, for her support and guidance throughout my time at UNCW. Her hard work, dedication, and attention to detail have helped me tremendously, and for that I am truly grateful. I would also like to thank my thesis committee, Drs. Raymond Pitts, Wendy Donlin Washington and Anne Hungerford for their encouragement and comments throughout the completion of my graduate degree. I would also like to thank animal care and everyone who helped to run my animals. v

6 DEDICATION I would like to dedicate this manuscript to my husband, Christopher Guido. His support and encouragement mean more to me than I could ever express. vi

7 LIST OF TABLES Table Page 1. Summary of the research findings for d-amphetamine s increasing or decreasing effects on schedule-controlled behavior and schedule induced behavior with shock and time-out The number of administrations of each dose to each subject Mean reinforcement rate (SR/min) in each component (RI 1-min & RI 6-min) during prepunishment, the first 10 days of punishment, and the last 10 days of punishment before saline was administered shown in Figure 1 for Pigeons 567, 863, and 838 with the ranges in parentheses Mean reinforcement rate including time spent in time-out (#R/(time-in duration + (# TO x TO duration))) during the first 10 days of punishment and the last 10 days of punishment shown in Figure 1 for Pigeons 567, 863, and Mean time-out rate (TO/min) during the first 10 days of punishment and the last 10 days of punishment before saline was administered shown in Figure 1 for Pigeons 567, 863, and 838 where the ranges are in parentheses Saline response rates in the RI 1-min and RI 6-min components for Pigeon s 567, 863, and 838 can be found in table 6 where ranges are in parentheses The mean time-out rate (TO/min) for Pigeon s 567, 863, and 838 at control, saline, and each dose of d-amphetamine (0.3, 1.0, 1.8, 3.0, and 5.6 mg/kg) are shown in table 7 where ranges are in parentheses Mean reinforcement rate (SR/min) for Pigeons 567, 863, and 838 at control, saline, and each dose of d-amphetamine (0.3, 1.0, 1.8, 3.0, and 5.6 mg/kg) in each component (RI 1-min and RI 6-min) where ranges are in parentheses Mean reinforcement rate including time spent in time-out (#R/(time-in duration + (# TO x TO duration))) for Pigeons 567, 863, and 838 at control, saline, and each dose of d-amphetamine (0.3, 1.0, 1.8, 3.0, and 5.6 mg/kg) vii

8 LIST OF FIGURES Figure Page 1. Overall response rate (R/min) across the last 10 days of the prepunishment baseline, the first 10 days of the punishment baseline, and the last 10 days before the first saline injection for Pigeons 567, 863, and 838 in the RI 1-min (open circles) and the RI 6-min (filled circles) components. Note that the y-axes are different ranges across pigeons and the gap in sessions for Pigeon 863 is because the punishment schedule was changed from RR3 to RR Mean response rate (R/min) as a percentage of saline rates as a function of the dose of d-amphetamine for Pigeons 567, 863, and 838 in the punished RI 1-min (open circles) and the unpunished RI 6-min (filled circles) components. The error bars are ranges. Note that the y-axes are different ranges across pigeons Mean response rates expressed as a percentage of saline rates as a function of the dose of d-amphetamine for Pigeons 567, 863, and 838 in the punished RI 1-min (open circles) and the unpunished RI 6-min (filled circles) components. The graphs on the left show the data from the first presentation of the punished component and first unpunished component in the session. The graphs on the right show the data from the second presentation of the punished component and second unpunished component in the session. The error bars are ranges...32 viii

9 d-amphetamine s Effects On Behavior Punished By Time-Out from Positive Reinforcement d-amphetamine According to the National Institute of Drug Abuse (NIDA, 2009) d-amphetamine is a stimulant and increases levels of dopamine in the brain. Dopamine is associated with movement, pleasure, and attention. Taken in large doses, d-amphetamine can cause euphoria and lead to abuse of the drug. Amphetamine has several other effects including rapid breathing and heartbeat, high blood pressure, loss of appetite, dilated pupils, increased focus, and decreased fatigue. When d-amphetamine is prescribed (e.g. for attention-deficit hyperactivity disorder (ADHD)), it is given as infrequently as possible, at the smallest dose required to achieve the desired effect of behavior change. If necessary, it is gradually increased until the desired effect is reached. Doses are individualized to the child and their needs, but the average dose is mg a day. Time-out and Stimulant Prevalence In the field of applied behavior analysis, time-out usually refers to the withdrawal of positive reinforcement, or the opportunity to earn positive reinforcement, contingent on behavior. Time-out can be non-exclusionary, in which the participant is not completely removed from the situation, or exclusionary, in which the participant is removed completely from the environment for a period of time contingent upon his/her unwanted behavior. Many teachers use the exclusionary technique by employing a separate room, a hallway (i.e., the student sits in the hallway outside of the classroom), or a partition (i.e., the student sits in the room, but his/her view is blocked). If the problem behavior is positively reinforced, then placing the child in exclusionary time-out

10 presumably removes the opportunity for the positive reinforcement and the problem behavior should decrease. Time-out is used often because of its ease of application, acceptability with the public, and its rapid suppression of behavior (Cooper, Heron, & Heward, 2007). In addition to time-out and other behavioral management techniques, prescription drugs have been used to manage children s behavior. Stimulant medication, such as Ritalin (methylphenidate) and Adderal (amphetamine) has been used to treat several behavioral disorders, including ADHD. Zuvekas, Vitiello, and Norquist (2006) found that Ritalin was prescribed to 2.7% and 2.9% of 0-18 year-old children in 1997 and 2002, respectively. The 2.9% in 2002 equals 2.2 million users. Use was highest with 6-12 yearolds (4.8% in 2002), next highest with year-olds (3.2%), and lowest with children under 6 years (0.3%). The use of these stimulant medications was more prevalent in males (4.0% in 2001) than in females (1.7%) and in the White population (3.6%) than in Black (2.2%) or Hispanic (1.4%) populations. Stimulants also are reported to be used illegally among children and adults. According to a 2009 NIDA study, 4.1% of 8 th graders, 7.1% of 10 th graders, and 6.6% of 12 th graders had abused amphetamine in the past year. NIDA also reported nonmedical use of Ritalin by 1.8 % of 8 th graders, 3.6 % of 10 th graders, and 2.1 % of 12 th graders. Nonmedical use of Adderall was reported by 2.0 % of 8 th graders, 5.7 % of 10 th graders, and 5.4 % of 12 th graders. Methamphetamine also has been abused, but this has been declining since 1999 by about two thirds to 1.2% in 8 th and 12 th graders and 1.5% in 10 th graders (Meyer & Serwach, 2008). Many children are prescribed stimulant drugs such as amphetamine, and many are using stimulants illegally. 2

11 Children using these stimulants are also experiencing time-out in school and, possibly, at home, which is why it is important to examine the effects of d-amphetamine on behavior punished by time-out in controlled laboratory settings. Stimulants and Punished Behavior When examining punished behavior in a laboratory setting, a procedure known as the conflict procedure (i.e., a conjoint schedule) has been developed by Geller and Seifter (1960). In this procedure, responding usually first is maintained by positive reinforcement, and the data obtained serve as the baseline. Then, punishment is added for the same response that produces reinforcement. Thus, a conflict is created as the same response produces both reinforcement and punishment. This schedule allows the experimenter to compare the before-punishment baseline data to the data once punishment is added. In many studies (e.g., Branch, Nicholson, & Dworkin, 1977), a two-component multiple schedule is used. The components alternate within the session, and each component has its own schedule of positive reinforcement for a designated amount of time. For example, in the Branch et al. study, two schedules were arranged: a random interval (RI) 6-min schedule of food presentation and an RI 1-min schedule of food presentation, and only one schedule was in effect at a time. Once responding is established in each component, punishment, whether it is shock or time-out, is added to one of the components (schedules) in the multiple schedule. Therefore, responding in the punished component results in food presentation and either shock or time-out, and responding in the unpunished component results only in food presentation. A benefit of using a multiple schedule is that the effects of other variables, such as drugs, can be 3

12 examined on both punished and unpunished behavior in the same subject within the same session. Shock Punishment and Stimulants In most of the research on punishment with nonhumans in laboratory settings, shock has been used as the punisher. For example, McMillan (1973) conducted a study to test the effects of different classes of drugs, including stimulants (d-amphetamine), on punished behavior in four pigeons. The pigeons pecking was maintained by a multiple fixed-interval (FI) 5-min FI 5-min schedule of food presentation. That is, in the presence of either a red or green keylight (that alternated), a peck after 5 min resulted in 4-s access to grain. Once this behavior was stable, punishment was added to one of the two FI 5-min components. In the punished component, each response resulted in electric shock (punisher). The intensity (2.5 ma to 5.2 ma) and duration (50-ms to 100-ms) were manipulated across sessions until there was a moderate suppression of responding for each pigeon. Ultimately, the rate of responding in the punished component was less than half the baseline rates in the unpunished component. McMillan (1973) then administered a range of doses of four classes of drugs: sedatives (chlordiazepoxide, diazepam, oxazepam, meprobamate, pentobarbital, and mescaline), major tranquilizers (chlorpromazine and tetrabenazine), narcotics (morphine, delta 9 THC and delta 8 THC), and stimulants (d-amphetamine and imipramine). The sedative class of drugs increased the rate of responding in the punished component at doses that did not increase rates in the unpunished component. Rates of responding were increased more by pentobarbital than any other sedative. The major tranquilizers decreased the rate of responding in both the punished and unpunished components. Delta 4

13 9 and delta 8 THC decreased responding in both the punished and unpunished components, as did imipramine at large doses. In contrast, morphine increased only the rates of punished responding. d-amphetamine only marginally increased rates of punished responding beyond control ranges and only at the 1.0 mg/kg dose. This increase was much less than the increase of unpunished responding, which was clearly increased by 0.3 and 1.0 mg/kg. Foree, Moretz, and McMillan (1973) performed a series of studies, as a follow-up to McMillan s study (1973), in which they examined the effects of different drugs on punished responding. In Experiment 1, three pigeons key pecked under a multiple FI 5- min fixed-ratio (FR) 30 schedule, with a limited hold (LH) of 60-s, of grain presentation. That is, in the presence of a red key light, the first key peck after 5 min resulted in 4-s access to grain, and in the presence of a blue key light the 30th key peck resulted in 4-s access to grain. If 30 responses did not occur within 60 s in the presence of the blue key, or if no response occurred within 60 s in the presence of the red key after the FI had elapsed, then the schedule components alternated without grain presentation. The range of baseline performance were responses/min (R/min) in the FI schedule without shock and were R/min in the FR schedule. d-amphetamine (0.3, 1.0, and 3.0 mg/kg) increased low rates of responding maintained by the FI schedule and decreased high rates maintained by the FR schedule when there was no punishment. During the punished condition, a 50-ms shock was presented according to an FR1 schedule during each component. The shock intensity ranged across phases of the experiment from 2.5 ma to 5.2 ma. At the higher shock intensities, response rates were almost completely eliminated, whereas at lower shock intensities the rates decreased to approximately 50% 5

14 in the FR and FI components compared to baseline rates. d-amphetamine decreased response rates at the 3.0 mg/kg dose in both components. In Experiment 2, Foree et al. (1973) examined the frequency of punishment. Four pigeons key pecked under a multiple FI 5-min, FI 5-min schedule of food presentation with a LH of 40 s. After a stable baseline was obtained, a shock-presentation schedule was added to both components. In the presence of the red key light, a shock was delivered for each key peck (FR 1), and in the presence of the green key light a shock of the same intensity and duration was delivered for every 30th key peck (FR 30). Response rates were lower by around 45% more in the FR 1 component than in the FR 30 component. Overall, response rates decreased as doses of d-amphetamine increased in both components. d-amphetamine s effects on responding punished by shock also have been examined in rats. In a study conducted by Evenden, Duncan, and Ko (1998), eight rats responded on an FI 40-s schedule. Then, the experimenters used a light to signal punishment, which was a 1-s, 0.6 ma electric shock delivered on an FR 20 schedule. Punished and unpunished responding was intermixed within a session; response rates decreased by up to 50% when punished compared to when unpunished. The effects of psychotomimetics and anxiolytics on punished and unpunished responding then were examined. The psychotomimetics they used were d-amphetamine and N-methyl-Daspartate antagonist (MK801), and a nonpsychomotor stimulant psychotomimetic 5- HT2A/C agonist, DOI. The anxiolytics they used were chlordiazepoxide, NS2710, pregabalin, citalopram, and yohimbine. 6

15 To examine the results, Evenden et al. (1998) divided each 40-s interval into 10, 4-s subintervals (bins). d-amphetamine increased punished and unpunished responses at each dose. In the unpunished component, 0.8 mg/kg and 1.6 mg/kg d-amphetamine tripled response rates in certain bins (Bins 5, 6, and 7). All doses increased responses in the punished component (by an average of 0.1 R/s which is a 15% increase from saline). There was also an increase in overall rates for both the punished and unpunished components especially at the 0.8 mg/kg and 1.6 mg/kg doses. There were decreased rates at the 1.6 mg/kg dose in the later bins of the sessions. Evenden et al. also found that the anxiolytics and psychotomimetics generally increased responding, although the increase was much more noticeable in the unpunished component than in the punished component. In general, Evenden et al. (1998) found that d-amphetamine slightly increased overall punished response rates with shock. Overall, McMillan (1973) and Evenden et al. (1998) found that d-amphetamine slightly increased behavior punished by shock, whereas Foree et al. (1973) found d-amphetamine decreased behavior punished by shock. Time-Out and Stimulants In a series of studies, McMillan (1967) compared the punishing effects of response-dependent shock and response-dependent time-out. The first set of studies was conducted to test the effects of shock and time-out on behavior. McMillan used squirrel monkeys that pressed a lever on a variable interval (VI) schedule of food presentation that was manipulated between subjects. He also manipulated the shock intensity. Responding was maintained under a multiple VI 1-min VI 1-min schedule of food presentation. In one of the components a 30-ms shock at 3.0 ma intensity was added, and 7

16 in the other component a 40-s time-out of total darkness was added. Each component was presented twice within a session. Response-produced shock and response-produced timeout suppressed the response rates to approximately the same degree, but there were some differences in the suppression. In the shock-punishment component, the monkeys responded more during the second presentation of the shock punishment, whereas in the time-out component their rates remained suppressed during repeated presentations. That is, the punishment effects of time-out were longer lasting than the punishment effects of shock. In the next study in this series of experiments, McMillan (1967) injected the monkeys with pentobarbital twice a week to see the effects on punished responding. He found that 1.0 mg/kg and 3.0 mg/kg pentobarbital increased response rates that were previously suppressed by both electric shock and time-out. Responses punished by shock were increased much more (about 8 times more) than those punished by time-out (50% increase). Overall, McMillan showed that shock and time-out produce similar effects as punishers and can therefore be used as a baseline for drug effects. Van Haaren and Anderson (1997) examined the effects of chlordiazepoxide (anxiolytic drug), buspirone (anti-anxiety agent), and cocaine on rats responding punished by time-out. The experimenters used six rats. Responding was maintained under a multiple RI 30 s RI 30 s schedule of food presentation. Then, in one of the RI 30-s components a 10-s time-out, in which all stimuli were extinguished, which was signaled by a Sonalert, was presented according to an RI 2 s schedule. The 5-min unpunished component was presented 3 times in a session, and the 7.5-min punished component was presented 2 times in a session. The rats were then injected with one of the three drugs. 8

17 Van Haaren and Anderson found that low doses of chlordiazepoxide (1.0 and 3.0 mg/kg) increased response rates in the unpunished component by about 10% for three of the rats and not by much for the other three rats. They also found that these doses increased response rates in the punished component by about 20% for two of the rats and only slightly for the other rats. There were dose-dependent decreases in the punished and unpunished component at the larger doses (10.0 and 30.0 mg/kg) for all of the rats. They also found that 1.7, 3.0, and 4.2 mg/kg buspirone decreased response rates in both components and that the doses 0.1, 0.3, and 1.0 mg/kg did not change response rates. They found that 17.0 and 30.0 mg/kg cocaine decreased rates in the unpunished component and that no dose changed response rates in the punished component. Overall, the results with cocaine were variable within and across rats and, therefore, inconclusive. d-amphetamine and Schedule-Induced Behavior Schedule-induced drinking (Falk, 1961) is said to occur when there is an increase in water intake based on the schedule of food presentation. In studies on scheduleinduced drinking, food and water first are available to the subject in its home cage. How much water the subject drinks under these conditions is measured during this baseline. Then the subject is placed in an experimental chamber, in which there is an intermittent schedule of food presentation in effect (e.g. a fixed time [FT] schedule in all of the following studies), and a water bottle is present. The amount of water drunk in the session is measured and compared to the amount of water that was drunk in the home cage when both food and water were constantly available. Schedule-induced drinking is said to occur when the amount of water drunk during experimental sessions is greater 9

18 than that drunk in the home cage. In addition, schedule-induced drinking tends to occur immediately after food delivery in the experimental chamber. d-amphetamine s effects on schedule-induced drinking have been examined with both shock and time-out as punishers. Pellon, Mas, and Blackman (1992) conducted a study to examine the effects of d-amphetamine and diazepam on punished and unpunished schedule-induced drinking in rats. They used six rats and a typical scheduleinduced drinking procedure as described above. In Phase 1 of the experiment, 60 food pellets were presented according to an FT 1-min schedule when a 100-ml water bottle was in the chamber. At the end of the session, they measured how many licks the rat made and how much the rat drank. The mean water intake for each rat ranged from 4 to 4.5-ml while they had free access to water; whereas, their schedule-induced water intake ranged from 17.2 to 26.0-ml. In Phase 2, Pellon et al. (1992) separated the rats into two groups so that rats were matched by the amount of schedule-induced drinking in Phase 1. The rats in the first group were administered d-amphetamine (0.25, 0.5, 1.0, and 2.0 mg/kg), and rats in the second group were administered diazepam (0.5, 1.0, 2.0 and 4.0 mg/kg). The administration of drugs slightly decreased the amount of water consumed in the chamber, but these slight changes in water consumption were within the range of variability of baseline and, therefore, cannot be considered reliable. After a return to baseline conditions, punishment was added, and the rats received food every minute if they did not lick the water spout. If they licked the water spout, there was a 10-s signaled delay (i.e., the houselight was turned off) until the next food presentation. If the rat licked the water spout during the delay, the delay reset to 10 s (i.e., 10

19 differential reinforcement of other behavior, DRO schedule was in effect). Pellon et al. (1992) found that the punishment significantly decreased the mean amount of water that the rats drank to about half of what they drank under the unpunished conditions. Under the punishment baseline, d-amphetamine increased punished schedule-induced drinking, but not unpunished schedule-induced drinking. The increase at the lower doses (0.25 and 0.5 mg/kg) for two of the rats was around 30% and within the range of control for the other rat. The 1.0 mg/kg dose increased the amount of water drunk threefold for one of the rats, increased it by almost double for another, and did not change it for the other. The licking rate was increased slightly in two of the rats and fell within the baseline range for the other at the 0.25 mg/kg dose. The licking rate for all of the rats was increased at the 0.5 mg/kg and 1.0 mg/kg doses. The 2.0 mg/kg dose decreased the licking rate for one of the rats, increased it by almost triple for another and increased it by 20 times the control rate for the other rat. In contrast, diazepam increased the amount of water drunk when the scheduleinduced drinking was unpunished and did not affect the punished schedule-induced drinking. Pellon et al. (1992) found that there was a dose-related effect, with small doses producing increases in water consumption in unpunished licking, whereas larger doses decreased water consumption and licking. Similar to what Pellon et al. (1992) found, Perez-Padilla and Pellon (2003) found that d-amphetamine increased water consumption that was reduced by negative punishment procedures. They used 24 rats and a typical schedule-induced drinking procedure. The 12 rats first were exposed to 30-min sessions with an FT 30-s schedule of food presentation. The rats were divided into six pairs based on their licking rate and the 11

20 amount of water consumed during baseline. Perez-Padilla and Pellon selected one rat from each pair to be the experimental rat for which every lick produced a responsedependent unsignalled 10-s delay to the next food pellet. The other rat in the pair was the control (yoked) rat and received the delay at the same time as its experimental counterpart regardless of its behavior (response-independent). The first group of 12 rats was the maintenance group in which the delay was added after polydipsia (excessive water intake) was induced, and the second group of 12 rats was the acquisition group, in which the delay was present from the outset of the experiment. After 30 sessions, the animals were injected with d-amphetamine ( mg/kg). In the experimental rats in the maintenance group, 1.0 mg/kg d-amphetamine increased licks per minute by 200% and water intake by 100%; the 3.0 mg/kg dose decreased rates. In contrast, d- amphetamine dose-dependently decreased licks per minute and water intake in the yokedcontrol rats. That is, the effects of d-amphetamine depended on the punishment contingency. In both the experimental and yoked-control rats in the acquisition group, d- amphetamine dose-dependently decreased licks per minute and water intake for both the experimental and yoked control rats. The 1.0 mg/kg dose decreased rates the most with the licking rate decreasing by almost 100% and the water intake decreasing by about 20%. The differential drug effects in the two groups of rats show that when punishment is introduced, after schedule-induced drinking was established (maintenance group) or during acquisition (acquisition group), affects the data. Perez-Padilla and Pellon (2006) examined the effects of d-amphetamine on punished and unpunished behavior at the same time. They used 16 rats and a typical schedule-induced drinking procedure to examine if the level of response suppression is an 12

21 important determinant of d-amphetamine s effects. The rats were divided into two groups by Perez-Padilla and Pellon to examine the level of response suppression. The first group was exposed to a multiple FT 30-s FT 45-s schedule, and the second group was exposed to a multiple FT 30-s FT 90-s schedule. A lick-contingent signaled delay was added to the FT 30-s component. Similar to Pellon et al. (1992), and Perez-Padilla and Pellon (2003), Perez-Padilla and Pellon (2006) also found that d-amphetamine ( mg/kg) dose-dependently increased (until the 3.0 mg/kg dose) punished schedule-induced drinking in both groups. However, there was also an increase in the unpunished rate (FT 90-s) at the 1.0 mg/kg dose. The baseline rates in the FT 45-s FT 90-s group, however, were very low. These data suggest that there may be an interaction between baseline levels of schedule-induced drinking and the effects of d-amphetamine, showing that the level of response suppression is an important determinant for the effects of d- amphetamine. d-amphetamine increased punished licking the most when it was the most reduced (FT 30s with delay and FT 90s). Therefore, the greater the decrease in the licking rate, the greater the effect of d-amphetamine. d-amphetamine has also been examined using shock as a punisher with scheduleinduced drinking with different results than those found with time-out as a punisher. Flores and Pellon (1998) found that d-amphetamine dose-dependently decreased the licks per minute with schedule-induced drinking punished by shock. In a subsequent study, Perez-Padilla and Pellon (2007) found within the same session that d-amphetamine decreased schedule-induced drinking when punished by shock and increased scheduleinduced drinking when punished by time-out. Therefore, when shock was used instead of 13

22 time-out to punish schedule-induced drinking, d-amphetamine decreased rates whether examined across phases or within the same session. Table 1 shows a summary of the research findings for d-amphetamine s increasing or decreasing effects on schedule-controlled behavior and schedule induced behavior with shock and time-out. Table 1 Summary of Findings Presented Shock Time-Out Increased Decreased Increased Decreased Schedule- Controlled Foree et al McMillan 1973 Evenden et al Schedule- Induced Flores, Pellon 1998 Perez-Padilla & Pellon 2007 Perez-Padilla & Pellon 2003 (Acquisition) Perez-Padilla & Pellon 2003 (Maintenance) Perez-Padilla & Pellon 2006 Pellon et al Perez-Padilla & Pellon

23 d-amphetamine and Schedule-Controlled Behavior Punished by Time-Out Different effects of d-amphetamine on rates of behavior have been found as a function of the type of punisher; that is, shock and time-out. d-amphetamine increased rates of schedule-induced drinking punished by time-out (Perez-Padilla & Pellon, 2007), but decreased rates of schedule-induced drinking punished by shock (Flores and Pellon, 1998). d-amphetamine generally decreased schedule-controlled response rates punished by shock (e.g., Foree et al., 1973). The effects of d-amphetamine on schedule-controlled behavior punished by time-out have not been examined. The present study replicated Branch et al. (1977) with time-out as the punisher and d-amphetamine as the drug. In their study, pigeons key pecking was maintained by a multiple RI 6-min RI 1-min schedule of mixed-grain presentation. Then a random ratio (RR) 3 schedule of 20-s time-outs was added to the RI 1-min component. In a subsequent phase, they punished responding in the RI 1-min component with shock presentation. Branch et al. showed that pentobarbital had different effects on schedule-controlled responding in pigeons dependent on the type of punisher. That is, response rates that were suppressed by time-out presentation were not increased by pentobarbital; in contrast response rates punished by shock were reliably increased (Branch et al., 1977). Similar to studies by Perez-Padilla and Pellon (2007), these results indicate that effects of drugs on punished behavior may depend on the type of punishment. In the present experiment, the effects of d-amphetamine on schedule-controlled response rates punished by time-out were examined. The procedure in the Branch et al. study was used so that effects of d- amphetamine could be examined on punished and unpunished behavior within the same session. 15

24 Current Experiment Responding was maintained in four pigeons using a multiple RI 6-min, RI 1-min schedule of food presentation with 20-s time-outs added to the RI 1-min component presented with a RR2 schedule for two of the pigeons and a RR3 schedule for the other two pigeons. The schedules were adjusted so that response rates in the punished component were approximately half of the previously unpunished rates. There were two components (RI 1-min and RI 6-min) with 3 presentations of each component, each 5 min long excluding time spent in time-out. There were no time-outs between components. The effects of d-amphetamine (0.3, 1.0, 1.8, 3.0, and 5.6 mg/kg) were examined. Based on the literature, d-amphetamine could decrease rates in the punished component and not decrease rates in the unpunished component. These differential effects also are predicted by data that show that d-amphetamine affects timing (Odum, Lieving, & Schall, 2002). That is, d-amphetamine would make it seem as though more time has passed while the pigeon is in time-out than has actually passed, therefore the time-out would seem longer and, therefore, the time-out would be more punishing. d-amphetamine could increase the rates of punished behavior and not change or decrease unpunished rates. This differential effect is predicted by a rate-dependency notion of amphetamine s effects. That is, the low punished rates are increased by d- amphetamine; whereas, the higher unpunished rates are not changed or decreased like the results found by Pellon, et al. (1992) and discussed by Odum, Lieving, and Schall (2002). These results would show that with schedule-controlled behavior, d-amphetamine produced effects dependent on type of punisher and consistent with the schedule-induced 16

25 literature results found by Pellon et al. (1992) and Perez-Padilla and Pellon (2007). Timeout would not seem as long and therefore would not be very effective in reducing the given behavior. Therefore, this also suggests that time-out may not be effective as a punisher if a child is taking a stimulant medication than if a child is not taking a stimulant medication. d-amphetamine could decrease both punished and unpunished response rates. This would show that d-amphetamine decreases schedule-controlled behavior with shock and with time-out, suggesting that time-out and shock are equitable punishers with schedule-controlled behavior like McMillan found in d-amphetamine could also increase both punished and unpunished response rates. This would show that d-amphetamine has different effects on schedule-controlled behavior based on the punisher suggesting that time-out and shock are not equitable punishers, like the results found by Perez-Padilla and Pellon in These results may reflect an overall stimulant effect in which behavior in general is increased. METHOD Subjects Four racing homer pigeons were used; one of which was experimentally naïve (Pigeon 190). The other three pigeons had responded under RI schedules and had a previous history of being injected with d-amphetamine, but had not received d- amphetamine for at least 6 months before the beginning of this study. Before training, the birds were reduced to 80% of their free-feeding weight. They were weighed each day before the session and again after the session and given the appropriate amount of Purina Checkers after the session to maintain 80% of their free feeding weight. Water and health 17

26 grit were continuously available in their individual home cages in a colony room in which humidity, light cycle (12 hr light/dark cycle, lights on at 7 a.m.), and temperature (69 to 71 degrees F) were controlled. Apparatus Four identical chambers were used in this experiment (BRS/LVE, Inc. model SEC-002). The chambers opened on the side; the keys were on the right side of the chamber when the door opened. The interior chamber was 35.0 cm by 30.5 cm by 36 cm. There were three, 2.5-cm diameter, response keys equidistant from one another on the right side of the chamber. The side keys were 9.0 cm from the corresponding side wall, and each key was 8.5 cm apart (center to center) from the next key in a horizontal line. The keys were 26 cm from the floor of the chamber. It took 0.25 N of force to be considered a key peck. The three keys could be transilluminated red, yellow, and green, though only the center key was used in this study. The food hopper, which was centered 11 cm below the center key, and the opening for the hopper was 5 cm by 6 cm. Three 1.2- W houselights were located 6.5 cm directly above the center key in a row. The houselights were red, white, and green; only the white light was used in this study. Each chamber had a ventilation fan that ran throughout the session and there was white noise white noise in the running room to mask any outside noises. The experiment was programmed and the data were collected by using Med Associates (Georgia, VT) interfacing and software on a Windows computer. The computer was in an adjacent room where interface equipment operated at 0.01-s resolution. 18

27 Behavioral Procedure Following magazine training, key pecking was shaped for 1 pigeon through differential reinforcement of successive approximations on the center key in the presence of a yellow keylight. The other three pigeons were already key pecking. Food presentation consisted of 3-s access to milo. The white houselight was illuminated whenever the keylight was on. A light illuminated the hopper when food was made available and the key lights and house light were no longer illuminated. Key pecking was then reinforced according to an FR 1 schedule for one session, each session in the presence of a red and a yellow keylight with 40 reinforcers. Then a multiple schedule of RI 10-s RI 10-s was introduced in which only one schedule was in effect at a time, and each was associated with either a red or yellow key light. Each component lasted for 10 s, and the other schedule was in effect for 10 s; each component was presented 3 times. There were no time-outs between components, and the only signal was the change of the key light to designate the different component. Over several days, the RI values and component lengths were gradually increased to a multiple RI 1-min RI 6-min schedule of food presentation with 2 components presented 3 times each within a session, each component lasted for 5 min. For Pigeons 838 and 863, the RI 1-min component was associated with the red key light, and the RI 6-min component was associated with the yellow key light. For Pigeons 190 and 567, the RI 1-min component was associated with the yellow light, and the RI 6-min component was associated with the red key light. One of the two components (RI 1-min or RI 6-min) was chosen at random to start the session every day. There was a 2-hr time limit in place for all six components, so if the session was not completed in 2 hr the session was terminated. 19

28 To determine whether the data were stable, the average response rates over the last 10 days were compared to the rates of the first 5 and last 5 days of those 10 days in each component. Each average from the 5-day blocks had to be within 10% of the average for the 10 days for response rates to be considered stable. Once response rates (R/min) were stable based on the criteria above under the baseline schedule, an RR schedule of time-out (houselight and keylight were extinguished and key pecks did not produce food) presentation was added to the RI 1-min component. Initially, the time-outs were 20 s in duration and followed responses immediately with a probability of.33. Because of lack of a clear punishment effect with Pigeons 863 and 190, the frequency of timeouts was increased. Therefore, the schedule of timeout presentation was an RR3 for Pigeons 838 and 567 and an RR2 for Pigeons 863 and 190. A peck that resulted in food presentation could not also produce a time-out. If food and a time-out are scheduled at the same time, the food is presented and the time-out is cancelled. The component length was the same during the non-time-out component as during the time-out component exclusive of the time spent in time-out (so time-in time was the same for the punished and unpunished components). Pharmacological Procedure d-amphetamine sulfate was dissolved in saline (0.9% sodium chloride) and injected 15 min prior to selected sessions. Injections were given into the breast muscle (i.m.) on alternating sides using a solution volume of 1.0 ml/kg. Injections were given no more than twice a week. Before the administration of drugs was given, two saline injections were administered and examined to ensure that there were no effects of the injection. If the response rates were outside of the control range when saline was 20

29 administered, then that pigeon was injected with saline again until there were two saline injections without a reaction. Then doses of 0.3, 1.0, 1.8 and 3.0 mg/kg were given two times each, and more determinations were given to certain pigeons based on the data. Each dose was given once before a dose was repeated. The 5.6 mg/kg dose was only given once to Pigeon s 863 and 567 because they stopped responding. Pigeon 863 had lasting effects from the dose where he did not respond the day following the injection as well as the day he was injected and ate no food for two days. The order of doses was random within each round of doses. Pigeon 190 never received a saline or drug administration. The number of administrations of each dose to each subject are presented in Table 2. Table 2 Number of administrations of saline and d-amphetamine. Pigeon Dose Saline mg/kg mg/kg mg/kg mg/kg mg/kg Data Analysis Only the first four components were used in data analysis because Pigeon 567 and Pigeon 863 regularly did not finish the session once punishment was added to the RI 1- min component. These two pigeons completed the first four components during baseline every day so those components were used for all subjects. Only three subject s data were used because Pigeon 190 showed no punishment effect. The time-out duration and 21

30 frequency were manipulated with this pigeon; however, timeout never punished his response rates. Because of this, he was never injected with d-amphetamine, and his data were not used in the current analyses. For the other three pigeons, the response rates were examined during the baseline period for both the RI 1-min and the RI 6-min components. That is, how often each pigeon responded in each of the components was calculated by dividing the number of responses by the time available for responding. Reinforcement time was not included in these calculations. Then the effect of time-out on these rates was examined. The time during punishment (time-outs) was not included in the denominator. The reinforcement rate also was calculated in both the RI 1-min and RI 6-min components, by dividing the number of reinforcers earned in a component by the time spent in the component, exclusive of time-out and reinforcement. Reinforcement rate also was calculated by dividing the number of reinforcers earned in a component by the time spent in the component inclusive of time-out #R/(time-in duration + (# TO x TO duration)). Timeout rate was calculated by taking the number of time-outs divided by time (minutes). The effects of d-amphetamine on response rates were examined by constructing dose-effect curves. Response rates were reported as percentage of saline. That is, the mean response rate following injections at a particular dose was divided by the mean response rate following saline injections. These calculations were conducted for all doses in each component. RESULTS Figure 1 shows that during the prepunishment baseline, response rates were higher in the RI 1-min component than in the RI 6-min component for each pigeon. Rates 22

31 were 30, 45, and 30 R/min higher in the RI 1-min component than in the RI 6-min component for Pigeons 567, 863, and 838, respectively, during the prepunishment phase. When timeouts were added to the RI 1-min component, there was a clear punishment effect for each pigeon. That is, response rates decreased by 58.40, 17.29, and R/min for Pigeons 567, 863, and 838, respectively. For Pigeons 838 and 567, response rates in the RI 6-min component did not change when punishment was added to the RI 1- min component. In contrast, for Pigeon 863, response rate in the RI 6-min component increased from to R/min, such that response rates in both components were approximately equal during the punishment baseline. For all of the pigeons, response rates in the first 10 days of punishment generally were similar to rates in the last 10 days of punishment before saline was administered. Pigeon 838 showed a decreasing trend in rates in the RI 1-min component, as did Pigeon 567, though less of a decrease. Both Pigeon 567 s and Pigeon 838 s rates in the RI 1-min component remained at the lower rates shown in the last 10 days of punishment (to the first 10 days of punishment) throughout the drug administration. Mean reinforcement rate (SR/min) in each component (RI 1-min & RI 6-min) during prepunishment, the first 10 days of punishment, and the last 10 days of punishment before saline was administered for Pigeons 567, 863, and 838 can be seen in Table 3 (ranges in parentheses). Mean reinforcement rate did not vary greatly from prepunishment to punishment. For Pigeon 567, reinforcement rate in the RI 1-min component increased slightly, and for Pigeon 863 reinforcement rate in the RI 6-min component almost doubled. For these two pigeons, however, the range of reinforcement rates across phases overlapped. 23

32 Responses/min RI 6 min RI 1 min Session Figure 1. Overall response rate (R/min) across the last 10 days of the prepunishment baseline, the first 10 days of the punishment baseline, and the last 10 days before the first saline injection for Pigeons 567, 863, and 838 in the RI 1-min (open circles) and the RI 6- min (filled circles) components. Note that the y-axes are different ranges across pigeons and the gap in sessions for Pigeon 863 is because the punishment schedule was changed from RR3 to RR2. 24

33 Table 3 Mean reinforcement rate (reinforcers/min of time in) with ranges in parentheses Prepunishment 1 st 10 punishment Last 10 punishment Pigeon RI 1-min RI 6-min RI 1-min RI 6-min RI-1 min RI-6 min ( ) ( ) ( ) 0.16 (0-0.40) 0.13 (0-0.30) 0.18 (0-0.30) 1.05 ( ) 0.92 ( ) 0.90 ( ) 0.26 ( ) 0.22 (0-0.60) 0.16 (0-0.40) 1.14 ( ) 1.02 ( ) 1.03 ( ) 0.13 (0-0.30) 0.25 ( ) 0.22 (0-0.50) Table 4 Reinforcement Rate (Reinforcers per minute including time-outs) Pigeon 1 st 10 punishment Last 10 punishment ( ) ( ) ( ) 0.13 ( ) 0.14 ( ) 0.19 ( ) Table 4 shows the mean reinforcement rate including time spent in time-out during the first 10 days of punishment and the last 10 days of punishment for Pigeons 567, 863, and 838. When reinforcement rate was calculated this way, the mean reinforcement rates in both components were more similar (compare Tables 3 and 4). Reinforcement rate decreased substantially (up to 90%) when time spent in time-out was included compared to when reinforcement rate was calculated without time spent in timeout for all three pigeons. 25

34 Table 5 shows the mean time-out rate (TO/min) during the first 10 days of punishment and the last 10 days of punishment before saline was administered for Pigeons 567, 863, and 838 (ranges in parentheses). Mean time-out rate stayed approximately the same for Pigeons 567 and 863 from the first 10 days of punishment to the last 10 days of punishment with only slight decreases. Time-out rate decreased from the first 10 days of punishment to the last 10 days of punishment the most for Pigeon 838. Table 5 Time-out Rate (number of timeouts per minute) Pigeon 1 st 10 punishment Last 10 punishment ( ) ( ) ( ) ( ) ( ) ( ) Figure 2 shows that for Pigeons 567 and 863, d-amphetamine tended to have dose-dependent decreases in response rates. d-amphetamine decreased response rates for Pigeon 567 at all doses in both components with no differential effects. For Pigeon 863, there were differential effects at the 1.0 mg/kg dose with increases in the unpunished component and decreases in the punished component. There were also slight differential effects at the 0.3 mg/kg dose with increases in both the punished and unpunished components, but with the unpunished component increasing more than the unpunished component. For Pigeon 838, there were differential effects at all doses with increases in the punished components at all doses (except 5.6 mg/kg), and decreases in the unpunished component at all doses. Saline response rates in the RI 1-min and RI 6-min 26

35 Percentage of Saline Rates Dose Effect Curves Unpunished Punished d -Amphetamine (mg/kg) Figure 2. Mean response rate (R/min) as a percentage of saline rates as a function of the dose of d-amphetamine for Pigeons 567, 863, and 838 in the punished RI 1-min (open circles) and the unpunished RI 6-min (filled circles) components. The error bars are ranges. Note that the y-axes are different ranges across pigeons. 27