|Home | About | Journals | Submit | Contact Us | Français|
Concurrent schedules of reinforcement are increasingly being used to investigate the reinforcing strength of abused drugs. A purported advantage of concurrent schedules is that the primary dependent measure, percent of responses emitted on the drug-associated manipulandum, is independent of the rate-altering effects of drugs. However, data supporting this hypothesis are rarely presented, which was one goal of the present study. In addition, we tested the hypothesis that drug-induced decreases in response rates provides an additional index to characterize abuse liability of drugs. The present study examined the relationship between response rate and response allocation (i.e., drug choice) when MDMA (0.03 – 0.3 mg/kg/inj) or cocaine (0.003 – 0.1 mg/kg/inj) was the alternative to food under a concurrent fixed-ratio (FR) reinforcement schedules in rhesus (n=4) and cynomolgus (n=16) monkeys, respectively. Increasing doses of MDMA or cocaine resulted in increased drug choice and dose-dependent decreases in overall response rates. For both drugs, response rates on the drug-associated lever were not affected by dose and were not different from saline. Furthermore, at all doses, rates of responding on the food-associated lever were significantly higher than response rates on the drug-associated lever. Finally, MDMA but not cocaine decreased food-reinforced responding, providing evidence for potential differences between the drugs. These results demonstrate that under concurrent food-drug reinforcement schedules, response rates on the drug-associated lever are independent of measures of reinforcement, while disruptions in food-maintained responding may be inversely related to abuse liability.
Concurrent and discrete-trial choice procedures are used to measure the relative reinforcing strength of abused drugs. When the alternative reinforcer to a drug injection is a non-drug stimulus, such as food, the procedure more closely resembles the clinical abuse condition (Katz, 1990), in which choosing drug results in forfeiting another reinforcer (i.e. food, money or job). Choice procedures have the purported advantage that the primary dependent measure, response allocation, is independent of the rate-decreasing effects of the self-administered drug (cf. Woolverton and Nader, 1990). Previous studies have suggested that response rates do not correlate significantly with drug choice or preference (Johanson and Aigner 1981; Woolverton and Johanson 1984; Gomez and Meisch, 2003). For example, in a discrete-trial choice paradigm, Woolverton and Johanson (1984) noted that when cocaine was preferred over d,l-cathinone (i.e., greater than 75% choice), response rates on the cocaine-associated lever were higher than rates on d,l-cathinone-associated lever in only 22% of those cases. The same is believed to be true when food-drug choice experiments are conducted (e.g., Aigner and Balster, 1979; Nader and Woolverton, 1991; Czoty et al., 2005; Negus, 2006; Banks et al., 2008). However, perhaps due to space limitations, supporting data are typically not presented.
One goal of the present study was to examine the relationship between response rates and drug choice under concurrent fixed-ratio (FR) schedules in monkeys. A second goal was to compare the relationships between food- and drug-associated response rates when responding was maintained by cocaine vs. 3,4-methylenedioxymethamphetamine (MDMA), two psychostimulants with high abuse liability. When studied under progressive-ratio (PR) schedules, which have also been described as measuring reinforcing strength (Hodos, 1961), cocaine generated significantly higher breaking points than MDMA (Lile et al., 2006) and under simple FR schedules of reinforcement, MDMA engendered lower response rates than cocaine (Beardsley et al., 1986; Lamb and Griffiths, 1987). However, when studied under food-drug choice conditions, both cocaine (Czoty et al., 2005) and MDMA (Banks et al., 2008) produce dose-dependent increases resulting in approximately 100% drug choice.
Examining response rates under concurrent schedules of reinforcement may provide information regarding differences in abuse potential reported under other schedules of reinforcement, especially if the behavioral profiles are different. Meisch and colleagues have proposed a measure “persistence of behavior” in which response rates are examined under concurrent schedules in which the ratio value is manipulated (Meisch, 2000; Gomez and Meisch, 2003). This analysis has been particularly informative with regard to comparing different drugs or different doses of the same drug. In the present study, we were interested in examining how food and drug response rates varied with dose and whether those behavioral profiles were unique between drugs. We hypothesized that when comparing doses of MDMA and cocaine that yield similar preference over food, the effects on food-maintained responding will be different, reflecting differences in abuse liability as reported using PR schedules. That is, another factor determining abuse liability is the direct effects of the reinforcing drug on other behaviors. In this manuscript, we present response rate data and relate them to previously published drug choice data (Czoty et al., 2005; Banks et al., 2008) in an attempt to test this hypothesis.
Adult male rhesus (Macaca mulatta; n=4) and cynomolgus (M. fascicularis; n=16) monkeys, each surgically prepared with an indwelling intravenous catheter and a subcutaneous vascular access port (Access Technologies, Skokie, IL; Czoty et al., 2005; Banks et al., 2008), served as subjects. Monkeys were weighed weekly and fed enough food daily (LabDiet #5038 Monkey Chow and fresh fruit) to maintain body weights at approximately 90-95% of free-feeding levels; subjects were fed at least 2 hours after completion of the experimental sessions. Water was freely available in the homecage. Rhesus monkeys were individually housed with visual and auditory contact with each other whereas cynomolgus monkeys were socially housed in groups of four as previously described (Czoty et al., 2005). For this study, data were collapsed across social ranks. All procedures were approved by the Animal Care and Use Committee (ACUC) of Wake Forest University and performed in accordance with the 2003 National Research Council Guidelines for the Care and Use of Mammals in Neuroscience and Behavioral Research. Environmental enrichment was provided as outlined in the ACUC of Wake Forest University Non-Human Primate Environmental Enrichment Plan.
Monkeys were trained under a concurrent FR (30 for rhesus; 50 for cynomolgus monkeys) schedule of food and drug (MDMA or cocaine; 0.003–0.3 mg/kg/inj) presentation (for details about the apparatus and training procedures see Czoty et al., 2005; Banks et al., 2008). Completion of the FR requirement resulted in the white lights being extinguished, the red lights being illuminated (for 30 sec in MDMA and 10 sec in cocaine studies) and the appropriate reinforcer being delivered. Responses emitted on the alternate lever before the completion of an FR reset the response requirement for the initial lever. Responding during illumination of the red lights had no scheduled consequences and was not counted in calculating response rates. The session ended after 30 total reinforcers had been earned or 60 min had elapsed. Each dose was available for at least five consecutive sessions and until responding was deemed stable (percent injection-lever responding ± 20% of the mean of three consecutive sessions with no trend). Doses were tested in mixed order (MDMA and cocaine studies) for each monkey.
Behavioral data were analyzed with repeated measures mixed linear regression using SAS Proc Mixed (Version 8.2, SAS, Cary, NC) with MDMA or cocaine dose as the factor. For this report, the primary dependent variables were response rates on the drug- and food-associated levers and overall response rates. Separate clocks were associated with each lever and began counting when the first response occurred on that lever; the clock associated with that lever stopped upon completion of the required ratio or with the first response on the alternate lever. Total responses on a particular lever were divided by the total time that lever was active (i.e. the subject was responding) and provided a measure of running response rates. Time out durations were not included in the running response rate calculations. Furthermore, another clock was used to determine session length and overall response rates were calculated with that value in the denominator and total responses on both levers in the numerator. Time out durations were included in the overall response rate calculations. When no responses occurred on a manipulandum during a behavioral session, a response rate of zero was recorded. This occurred infrequently, primarily on the last days each dose was available, and only on the non-preferred lever. For data related to the frequency of reinforcement, we refer the reader to our previous publications (Czoty et al., 2005; Banks et al., 2008). Tests of the main effects of MDMA or cocaine dose on food- or drug-maintained or overall response rates were performed using repeated measures analysis of variance (ANOVA). In the presence of a main effect, a Tukey HSD post-hoc test was conducted. Tests for differences between food- and drug-maintained responding across doses were performed using a paired t-test. Differences were considered significant at the 95% level of confidence (p<0.05).
As reported previously (Czoty et al., 2005), cocaine choice increased as a function of dose with the 0.03 and 0.1 mg/kg/inj doses being chosen over food on approximately 90% of the trials (Table 1). When overall response rates were analyzed, there was a significant main effect of cocaine dose (F4,53 = 26.95, p < 0.001; Fig. 1A, open diamonds). Post-hoc analysis demonstrated a significant decrease (p < 0.05) in overall response rate at the 0.03 and 0.1 mg/kg/inj cocaine doses compared to the 0.003 mg/kg/inj dose. When saline was the alternative to food, the mean (SEM) running response rate on the food-associated lever was 4.07 (0.51) responses/sec (Fig. 1A, filled circles); this rate did not change significantly when the alternative was cocaine (0.003-0.1 mg/kg). The dose of cocaine that resulted in approximately equal preference with food (0.01 mg/kg) was associated with approximately equal cocaine (3.10 ± 0.62 responses/sec) and food (3.48 ± 0.55 responses/sec) running rates. Running rates on the cocaine-associated lever did not change significantly as a function of cocaine dose and were not different from running response rates when saline was available (Fig. 1A, filled symbols).
As reported previously (Banks et al., 2008), MDMA choice increased as a function of dose, with the 0.1 and 0.3 mg/kg/inj doses being chosen over food on approximately 90% of the trials (Table 1). When overall response rates were analyzed, there was a significant main effect of MDMA dose (F3,9= 21.7, p < 0.001; Fig. 1B). Post-hoc analysis demonstrated a significant (p < 0.001) decrease in overall response rates at the 0.1 and 0.3 mg/kg/inj MDMA doses compared to 0.03 mg/kg/inj dose. When saline was the alternative to food, the mean (SEM) running response rate on the food-associated lever was 2.25 (0.56) responses/sec; this rate changed significantly as a function of MDMA dose (F3,9 = 5.33, p < 0.05; Fig. 1B, filled circles). Post-hoc analysis demonstrated that response rates on the food-associated lever were significantly lower when 0.3 mg/kg/inj MDMA was available compared to when 0.03 mg/kg/inj MDMA was the alternative to food. Running rates on the MDMA-associated lever did not significantly change as a function of MDMA dose and were not different from running response rates when saline was available (Fig. 1B, filled squares).
The main goal of the present study was to test the hypothesis that measures of response rates did not vary in a manner consistent with drug choice under conditions in which the alternative was food reinforcement. A second goal was to compare drug choice and response rates for two drugs, cocaine and MDMA, as a function of dose. Irrespective of whether cocaine or MDMA was the alternative to food, increases in drug dose resulted in increases in frequency of drug choice and decreases in overall response rates. Drug-associated running response rates (i.e., responses on each lever divided by time on each lever) for cocaine and MDMA did not change as a function of dose and were not different from running rates when saline was studied. When choice was between food and cocaine, food-associated response rates did not significantly change with increases in cocaine dose. In contrast, increases in MDMA dose resulted in decreases in food-associated response rates. The findings that increases in MDMA choice, but not cocaine choice, occurred at doses that disrupted non-drug reinforced responding may reflect the differences between cocaine and MDMA in other measures of reinforcing strength such as break points under PR schedules (Lile et al., 2006).
Previous studies involving drug-drug choice paradigms in which responding was maintained under FR schedules of reinforcement have suggested that response rates do not correlate with drug choice (Johanson and Aigner 1981; Woolverton and Johanson 1984; Gomez and Meisch, 2003). Although similar findings have been suggested in studies involving food-drug choice, response rate data have not been presented to support those observations. In the present study, since food-associated response rates were typically higher than drug-associated rates, the only occasions in which higher response rates were associated with the preferred choice was at low cocaine or MDMA doses in which food was preferred to drug. These findings reinforce the observations that under concurrent-access conditions the effects of drugs on response rate are independent of reinforcer strength (Woolverton and Nader, 1990).
The dose-dependent increases in drug choice (Table 1; data from Czoty et al., 2005; Banks et al., 2008) for both cocaine and MDMA resulted in dose-dependent reductions in overall response rates. This is likely a result of the dose-dependent increase in session length (data not shown) that occurred under both concurrent cocaine-food and MDMA-food schedules of reinforcement. In fact, at the highest doses tested, subjects did not typically earn all reinforcers and thus the session did not terminate until 60 minutes had elapsed. Previous studies using FR reinforcement schedules have shown that increases in self-administered drug doses resulted in increases in post-reinforcer pause times (Pickens and Thompson 1968; Dougherty and Pickens 1973) as well as decreases in response rates as the session progressed (Goldberg 1973). While post-reinforcement pause time was not recorded in the present studies, an increase in pause time would have increased session length and decreased overall response rates.
Unlike overall rates, running response rates on the cocaine- or MDMA-associated levers did not vary according to the available dose. In fact, running response rates on the drug-associated lever were not significantly different from saline at any dose. While increasing the dose of either drug would eventually produce decreases in running rates on the drug-associated levers, it is important to note that we were studying dose ranges that yielded nearly complete food preference up to cocaine and MDMA doses that were chosen on nearly every completed trial. Perhaps of even greater interest were the dose-dependent changes in food-maintained responding. Increases in MDMA dose resulted in dose-dependent decreases in response rates on the food-associated lever, while increases in cocaine dose did not significantly affect response rates on the food-associated lever. These findings suggest different behavioral mechanisms mediating changes in drug choice for cocaine and MDMA. The observation that changes in cocaine preference occur in the absence of disruption in non-drug reinforced responding, while MDMA preference occurs at doses that disrupt food-maintained responding, may provide a novel indication of the greater abuse potential of cocaine compared to MDMA.
The authors thank Susan Nader and Tonya Calhoun for assistance through out these studies. No authors have financial conflicts of interest with the research described in this manuscript. The research was supported by the National Institute on Drug Abuse grants DA10584, DA06634 and DA020281.
This research was supported by National Institute on Drug Abuse grants DA-10584, DA-06634, and DA-020281.