History of Drug Exposure as a Determinant of Drug Self-Administration


The purpose of this paper is to review how a drug’s effectiveness in initiating and maintaining self-administration can be influenced by a subject’s past experience with drugs. Drug self-administration by humans and laboratory animals is considered an instance of operant behavior (), controlled by the subject’s genetic constitution, past history, and the current circumstances of drug availability (of Skinner, 1938). The influence of history of drug exposure on current drug-maintained behavior may be controlled, in turn, by the particular drugs and doses employed and the conditions under which the drug is administered. This discussion will focus on the ways in which a history of drug exposure can control later drug self-administration in laboratory animals.

Effects of history of drug exposure on initiation of drug self-administration

In order to study drug self-administration by laboratory animals, an experimenter must set up a situation in which subjects are exposed to some contingency between the occurrence of a specific response and delivery of a particular drug. For many drugs, no explicit behavioral or pharmacologioal history is necessary for the drug to maintain behavior. In one initial study, for example, Deneau et al. () surgically prepared drug-naive rhesus monkeys with Indwelling venous catheters and presented the monkeys with a response lever. Presses on the lever delivered an intravenous injection of a drug. If a monkey did not press the lever at all during the experimental periods, a raisin or bit of candy was taped to the lever so that the monkey would depress the lever when grabbing for the food. Under these conditions, for the majority of monkeys tested, lever pressing was Initiated and maintained by Injection of appropriate doses of morphine, codeine, cocaine, d-amphetamine, pentobarbital, or ethanol. On the other hand, lever pressing was not maintained by injections of nalorphine, chlorpromazine, mescaline, or saline. These initial results have been amply replicated and extended to other drugs by numerous investigators (). Thus, for many drugs, a history of drug exposure is not necessary for the drugs to function as reinforcers. Exposure to the contingency between a specific behavior and drug delivery is sufficient for the drugs to function as reinforcers and maintain subsequent behavior.

While prior drug administration is not necessary for the Initiation and maintenance of self-administration of many drugs, it can hasten the development of asymptotic performance at a particular drug dose and schedule parameter. For example, if a monkey whose behavior has been maintained by intravenous injection of codeine loses its catheter and does not self-administer the drug for some period of time, replacement of the catheter is quickly followed by a return to the previous response rates. Additionally, exposure to schedule contingenoies for the delivery of other drugs or other events such as food can increase rates of behavior that may be initially low when maintained by a drug such as ethanol. Winger and Woods () reported that for certain rhesus monkeys, intravenous delivery of ethanol under a FR 1 schedule maintained fen responses when available 24 hr per day. When injections of cocaine or methohexital replaced the ethanol, responding was initiated and maintained during 3 hr access periods. When ethanol was then reintroduced, responding continued at the higher levels. Moreover, the intake of ethanol for these subjects under the 3 hr access conditions did not differ from that for subjects that initiated ethanol self-administration without exposure to cocaine or methohexital.

Under certain conditions, the availability of a drug is not sufficient for it to function as a reinforcer. In a well-studied example, behavior often is not readily maintained by the oral delivery of drugs. However, a history of drug exposure that ensures that an organism will readily ingest an effective dose of a drug can increase the likelihood that certain drugs will serve as oral reinforcers. A good example of this effect of history of drug exposure is provided by the procedures developed by Meisch and colleagues () to establish ethanol as an oral reinforcer in rhesus monkeys. When ethanol is presented to monkeys, they do not readily drink large quantities, and, above small concentrations (5%), they may drink water to the exclusion of ethanol (). Under certain schedules of food delivery (), however, monkeys will adjunctively drink large quantities of ethanol (). After a history of adjunctive or schedule-induced drinking of gradually increasing concentrations of ethanol, high concentrations of ethanol can maintain responding in the absence of the original inducing schedule (). Subsequent work () has shown that a variety of inducing schedules are also effective in establishing drugs such as etonitazene and phencyolidine as oral reinforcers in rhesus monkeys. These compounds do not maintain behavior initially, but after gradually Increasing concentrations have been consumed under an inducing procedure, each will maintain behavior in the absence of the original Inducing condition. To date, the particular history used to establish a drug as a reinforcer has not been shown to control the later behavior maintained by the drug. The behavior maintained by oral ethanol after exposure to an inducing schedule, for example, varies as a function of the current schedule conditions in the same nay as does behavior maintained by the intravenously delivered drugs that do not require prior exposure to inducing schedules to function as reinforcers ().

Importance of physiological dependence

For those drugs that have been extensively studied, it appears that prior physiological dependence is not necessary for a drug to function as a reinforcer (). The conditions under which such dependence is maintained, however, can influence the later probability of drug self-administration. In the case of physiological dependence on morphine, the likelihood that a post-dependent subject will self-administer morphine is controlled, in part, by the conditions under which the dependence was maintained. Rats that have maintained their physiological dependence on morphine by oral or intravenous self-administration will self-administer more morphine following a withdrawal period than will subjects that received the same maintenance doses of morphine noncontingently ().

Current physiological dependence can alter the likelihood that certain drugs will serve as positive reinforcers. In particular, narcotic dependence can alter the reinforcing properties of a variety of narcotic mixed agonist-antagonists. While morphine-like agonists such as morphine, heroin, and the systemically active met-enkephalin analogue FK 33-824 maintain behavior in both nondependent and morphine-dependent rhesus monkeys (, mixed agonist-antagonists such as profadol, propiram, and pentazooine maintain behavior only in nondependent monkeys (see review by Hoffmelster, 1979). A second group of mixed agonist-antagonists, including nalorphine and cyclazocine, and the narcotic antagonist naloxone generally do not maintain responding by either nondependent or morphine-dependent monkeys (). Morphine dependence can also alter the negative reinforcing properties of the mixed agonist-antagonists and antagonists. The mixed agonist-antagonist profadol maintains responses leading to the termination or postponement of its injection in morphine-dependent monkeys, but not in nondependent monkeys. The mixed agonist-antagonists nalorphine and cyclazocine and the antagonist naloxone, on the other hand, maintain responding leading to termination or postponement of their injection in both dependent and nondependent monkeys (). The doses of these drugs required to maintain such behavior, however, are up to 1000 fold lower in dependent than in nondependent monkeys.

The reinforcing properties of narcotic antagonists can be altered dramatically under certain conditions (). As described above, the narcotic antagonist naloxone will readily function as a negative reinforcer, maintaining behavior leading to the postponement or termination of its injection in both dependent and nondependent monkeys (). In morphine-dependent monkeys with an appropriate behavioral history, however, the same naloxone doses that maintain behavior leading to postponement or termination of their injection will also maintain behavior leading to the presentation of an injection. Downs and Woods () conditioned morphine-dependent monkeys to terminate and/or postpone injections of 2 microgram/kg naloxone. Characteristic fixed-ratio performance was maintained by termination and postponement of naloxone injections. Then, the schedule oontingencies were changed so that completion of each ratio produced a brief light flash; completion of every fifth or tenth ratio produced the light flash and an injection of naloxone. Behavior was maintained by the injection of naloxone in these morphine-dependent monkeys for as many as fifteen sessions. This apparently disparate effect of a presumably noxious pharmacological stimulus underlines the importance of the behavioral contingencies under which a subject is exposed to a drug in determining the later likelihood that the drug will maintain behavior leading to its administration.

Effects of self-administration history

A history of drug self-administration can influence the dose of a drug that will subsequently maintain behavior. In general, behavior is maintained by lower doses of drug in subjects with an extensive self-administration history than in subjects with a more limited history. For example, Goldberg () showed that a low cocaine injection dose (12 microgram/kg) initially failed to maintain fixed-ratio responding in monkeys with a limited history of cocaine-maintained behavior, but maintained high response rates in the same subjects after a period during which responding was maintained by higher cocaine doses. The actual response rates maintained by certain doses of a drug can also be altered by a history of drug-maintained behavior. For example, Downs and Woods () reported that the response rates maintained by injections of low doses of cocaine (3 and 10 miorogram/kg) in rhesus monkeys increased dramatically when these doses were retested after exposure to other cocaine doses. Similarly, Carney et al. () showed that the response rates maintained by several doses of ethanol increased when monkeys had a history of behavior maintained by higher ethanol doses.

The rate and pattern of behavior maintained by one drug can also influence both the initial pattern of intake of a substituted drug and the dose of that drug that will maintain behavior. For example, Schlichting et al. () reported that when amphetamine was substituted for cocaine, codeine, or pentobarbital under a fixed-ratio schedule in rhesus monkeys, the pattern of behavior initially maintained by amphetamine varied with the drug used to engender responding. Initially, the spacing of amphetamine injections was similar to that maintained by the maintenance drug. Cocaine maintained responses at regular intervals throughout experimental sessions, and all substituted amphetamine doses (0.005 to 0.05 mg/kg) maintained response rates above those maintained by saline, with responses spaced at regular intervals. The maintenance drugs codeine and pentobarbital, on the other hand, maintained frequent Injections at the beginning of the session, followed by long pauses interspersed with bursts of injections over the remainder of the session. When substituted for these drugs, low amphetamine doses (0.005 and 0.01 mg/kg) maintained patterns of responding similar to those maintained by codeine or pentobarbital. When the high dose of amphetamine (0.05 mg/kg) was substituted for these drugs, however, several injections were taken rapidly, followed by no responding until the end of the experimental session. Thus, as a result of the pattern of injections engendered by the maintenance drug, higher doses of amphetamine maintained more behavior when substituted for cocaine than when substituted for codeine or pentobarbital.

The drug used to maintain behavior in the monkey can also alter the behavior maintained by substitutions of narcotic agonists and mixed agonist-antagonists. Hoffmeister and Schlichting () reported that codeine, morphine, & propoxyphene, pentazocine, and propiram will maintain behavior at lower doses when substituted for codeine than when substituted for cocaine. In addition, although the doses of each narcotic that maintained the maximal number of Injections did not vary with the drug used to engender responding, the maximally effective doses of all the narcotics except morphine maintained more injections when substituted for codeine than when substituted for cocaine. As was the case for amphetamine, such differences in behavior may have resulted from the different patterns of drug Injections engendered by the maintenance drugs. Thus, under similar behavioral schedules, the asymptotic pattern of drug intake can vary markedly among drugs from different pharmacological classes. When behavior is initially established with a particular drug, however, the pattern of intake maintained by that drug can control the initial pattern of intake of quite different drugs.

Under certain conditions, a monkey’s self-administration history can alter not only the dose of a substituted drug that will maintain responding and the initial pattern of such responding, but also the likelihood that any dose of the new drug will maintain behavior. For example, self-administration of the antitussive dextrorphan is controlled, in part, by the subject’s self-administration history. Dextrorphan does not maintain behavior following one type of self-administration history, but readily maintains behavior following certain other histories. When dextrorphan is substituted for codeine under a FR 30 TO 10 min schedule of intravenous injection in rhesus monkeys, no dose maintains response rates higher than those maintained by saline  (). However, when dextrorphan is substituted for codeine under a FR 1 schedule, it maintains response and injection rates similar to those maintained by codeine. Figure 1 compares the behavior maintained by dextrorphan when substituted for codeine under these two conditions. The upper panel shows the response rates and injections per hour maintained under the FR 1 (or CRF) schedule by codeine (C), saline (S), and various doses of dextrorphan. Dextrorphan injection doses of 0.32 and 0.56 mg/kg maintained response rates similar to those maintained by 0.1 mg/kg codeine. However, as shown by the closed circles in the lower panel, the same doses of dextrorphan did not maintain responding when substituted for 0.32 mg/kg codeine under a different schedule, FR 30 TO 10 min.

These differences in the ability of dextrorphan to maintain behavior when substituted for codeine may be due to several factors. Dextrorphan may be relatively ineffective in maintaining behavior at high ratio requirements () . Alternatively, the monkeys’ behavioral histories may have contributed to the differences in dextrorphan self-administration. The monkeys performing under the FR 1 schedule continued to respond for an average of 20 injections per session when saline replaced codeine, while the monkeys performing under the FR 30 TO 10 min schedule rarely completed the ratio requirement to deliver more than 4 or 5 injections. The maintenance of behavior under the FR 1 contingenoies may have resulted in the monkeys’ self-administering sufficient dextrorphan for it to acquire a reinforcing function.

The importance of self-administration history in controlling the ability of dextrorphan to maintain behavior under FR schedules was given added weight by the results of an additional experiment (). In this study, other monkeys self-administered the dissociative anesthetic ketamine under the FR 30 TO 10 min schedule of intravenous injection. During selected sessions, various doses of dextrorphan were substituted for the ketamine maintenance dose. As shown by the open circles in the lower panel of Figure 1, dextrorphan readily maintained behavior when substituted for ketamine under the FR 30 TO 10 min schedule. Thus, under the FR 30 TO 10 min schedule, dextrorphan maintained response rates higher than those maintained by saline when substituted for ketamine but not when substituted for codeine. Two other compounds, the dissociative anesthetic phenoyolidine and the analgesic dexoxadrol, also maintained behavior under the FR 30 TO 10 min schedule when substituted for ketamine but did not maintain behavior when substituted for codeine ().

These differences in the ability of dextrorphan, dexoxadrol, and phencyclldine to maintain behavior when substituted for ketamine or for codeine may be controlled in part by the similarities in the behavioral properties of these compounds and ketamine. These drugs that maintained behavior only when substituted for ketamine share discriminative stimulus effects with ketamine, but not with codeine, in the rhesus monkey (). Common discriminative effects among ketamine, phencyclidine, dextrorphan, and dexoxadrol have also been reported in rats, pigeons, and squirrel monkeys (). Similarities among the discriminative stimulus properties of ketamine and those of phencyclidine, dexoxadrol, and dextrorphan may increase the relnforcing effectiveness of the latter three compounds when substituted for ketamine as compared to their effectiveness when substituted for codeine.

The control of the reinforcing effectiveness of a substitution drug by the maintenance drug Itself Is modulated by several factors. The maintenance drug is not a primary determinant of the ability of certain drugs to maintain behavior. For example, codeine, cocaine, and ketamine each maintain behavior when substituted for the other, while certain other drugs, such as cyclazocine and SKF-10,047 (N-allyl-normetazocine), do not maintain behavior when substituted for either codeine or ketamine (). The duration of exposure to a substitution drug and the prevailing schedule contingencies may also modulate the effects of the original maintenance drug. For example, with repeated exposure to phencyclidine or dextrorphan these compounds will maintain behavior in monkeys whose behavior is initlally maintained by codeine or cocaine (). Furthermore, drug-naive monkeys will initiate and continue responding leading to phencyolidine lnjection under FR 1 schedules during repeated daily sessions (). Thus, control of the self-administration of a new drug by a subject’s self-administration history varies with the particular drug under study, the drug with which the subject is experienced, and the behavioral conditions under which the new drug and the maintenance drug are available.

Effects of prior pairing between environmental stimuli and drug administration on self-administration behavior

The general environmental context in which prior drug administration has occurred can be an Important determinant of drug self-administration. For example, environmental stimuli paired with morphine self-administration can control the degree of self-administration by subjects previously dependent on morphine. If subjects self-administer sufficient morphine to develop physiological dependence in one environment and are subsequently withdrawn and then reexposed to morphine, the probability that they will later self-administer morphine varies as a function of the similarity of the environments in which the initial self-administration and reacquisition occur (). Rats exposed to the same environment in which self-administration originally occurred drink much more morphine after withdrawal than do rats reexposed to morphine in a different environment after withdrawal. Thus, the environmental stimuli associated with previous narcotic self-administration can control the likelihood that morphine self-administration will be reestablished in post-dependent subjects.

Under appropriate circumstances, environmental stimuli paired with the scheduled delivery of a drug can powerfully control the rate and pattern of ongoing drug-reinforced behavior (). Following exposure to certain behavioral schedules, stimuli paired with the administration of drugs such as cocaine and morphine can control behavior in the same way as do injections of the drug themselves. Moreover, the environmental stimuli associated with prior nonoontingent administration of one drug can control the later self-administration of a second drug. For example, under appropriate conditions, stimuli associated with narcotic antagonists can produce conditioned changes in the rate of morphine self-administration by morphine-dependent subjects. In morphine-dependent monkeys, administration of the antagonist nalorphine increases the rate of responding maintained by morphine (). With a history of repeated exposure to nalorphine, these increases occur with a much shorter latency and can be elicited by environmental stimuli paired with nalorphine (). Such conditioned stimuli can produce large but transitory increases. In morphine self-administration in morphine-dependent subjects. These conditioned stimuli are also capable of eliciting certain of the signs of morphine withdrawal, including emesis, salivation, changes in heart rate, and disruption of the rate of food-maintained operants (). These latter oonditioned stimulus effects, in contrast to the effects on morphine self-administration, are remarkably resistant to extinction and persist after monkeys have been withdrawn from morphine for two to four months.

A history of exposure to nalorphine can also control its potency in altering rates of morphine self-administration. Goldberg et al. () assessed the effects of nalorphine on the rate of responding maintained by morphine injections in morphine-dependent monkeys. In monkeys with a limited history of nalorphine injections, high nalorphine doses (1 to 3 mg/kg) suppressed the rate of morphine-maintained responding. In contrast, in monkeys that had received ascending doses of nalorphine and so had several sessions’ experience with lower nalorphine doses, 1 and 3 mg/kg nalorphine slightly increased the rate of morphine self-adminlstration. Likewise, in monkeys repeatedly exposed to 0.1 mg/kg nalorphine injections, the first injection of 1 mg/kg nalorphine increased the rate of morphine self-administration in two of four monkeys. This increase was transitory, however; the second injection of 1 mg/kg nalorphine did not Increase morphine self-administration, and all succeeding 1 mg/kg nalorphine Injections markedly suppressed morphine-maintained responding. It is likely that, with repeated exposure to low doses of nalorphine, interoceptive stimuli associated with the injection procedure became conditioned stimuli for increases in morphine self-administration. The Initial effect of the higher nalorphine dose was then a conditioned increase in responses maintained by morphine. With repeated exposure to high nalorphine doses, this response rapidly extinguished.

Recently, Herling () has presented evidence that a history of exposure to the narcotic antagonist naltrexone may also produce conditioned changes in the rate of narcotic-maintained behavior in nondependent monkeys. In these experiments, responding was maintained by codeine or food in alternate components of a multiple reinforcement schedule. Low doses of naltrexone antagonized the actions of codeine, increasing the injection dose of codeine required to maintain behavior and the cumulative dose necessary to decrease rates of food-maintained behavior. Higher doses of naltrexone suppressed responding maintained by all doses of codeine. In certain monkeys, some doses of naltrexone initially increased the rates of behavior maintained by codeine but suppressed behavior following repeated exposure. This suppression of codeine-maintained behavior by naltrexone was often greater than the effect produced by substituting saline in the session; i.e., extinction (). Herling and others () have suggested that such decreases in narcotic-maintained behavior may reflect a punishing effect of agonist-antagonist combinations, an effect that may be exacerbated by a history of repeated exposure to such combinations.


Drug self-administration is controlled, in part, by the subject’s history of drug exposure. Although a history of drug administration is not necessary for many drugs to function as reinforcers, prior exposure can increase the likelihood that certain drugs, such as ethanol, will maintain behavior. While it has been demonstrated that physiological dependence is not necessary for a drug to function as a reinforcer, the conditions under which such dependence is maintained can control the later self-administration of the drug. Once drug-maintained behaviors are established, the particular drug that maintains behavior can influence the Initial pattern of intake of a new drug and thus the dose of that drug that will maintain behavior. Additionally, under certain conditions, similarity between the discriminative stimulus effects of the drug that previously maintained behavior and those of a new drug can increase the likelihood that the new drug will function as a reinforcer. Finally, stimuli that have been paired with drug administration can powerfully control later drug-maintained behavior, the direction of such control being determined by the conditions under which such pairing occurred. In summary, both the type of drug with which a subject has experience as well as the contingencies governing that experience contribute to subsequent drug self-administration.


Selections from the book: “Behavioral Pharmacology of Human Drug Dependence”. Travis Thompson, Ph.D., and Chris E. Johanson, Ph.D., eds. Presents a growing body of data, systematically derived, on the behavioral mechanisms involved in use and abuse of drugs. National Institute on Drug Abuse Research Monograph 37, July 1981.