An Addiction Science Network Resource

Reprinted from M.A. Bozarth (1987), Intracranial self-administration procedures for the assessment of drug reinforcement. In M.A. Bozarth (Ed.), Methods of assessing the reinforcing properties of abused drugs (173-187). New York: Springer-Verlag.
Brain Reward System
Back to book TOC
(Search book contents) (Search entire ASNet)


Chapter 9

Intracranial Self-Administration Procedures
for the Assessment of Drug Reinforcement

Michael A. Bozarth
Center for Studies in Behavioral Neurobiology
Department of Psychology
Concordia University
Montreal, Quebec, Canada H3G 1M8


Procedures used for intracranial drug self-administration are described. The experimental approach uses lever-pressing and is a direct extension of methods used in intravenous drug self- administration studies. Some of the important factors that must be considered when conducting these experiments are discussed. Although intracranial self-administration potentially offers several advantages for the study of drug reinforcement, methodological difficulties have severely limited the routine application of this technique. 



The general procedures that have been developed for intravenous self-administration can be adapted to study reinforcement from drug injections directly into the brain. These injections may be into the cerebral ventricles (i.e., intraventricular) or directly into brain tissue (i.e., intracranial). The operant response requirements are usually the same as those used with intravenous drug self-administration (i.e., lever pressing). Many of the control procedures necessary to assure the validity of the conclusions drawn from these studies are also comparable, although several new types of controls must also be tested during intracranial self-administration.

There are a number of potential advantages to studying reinforcement from microinjections directly into the brain. First, injection sites are proximal (at least in theory) to their site of action; this minimizes the influence of diffusion barriers such as the blood-brain barrier and circumvents both first pass and local metabolism of the test compound that can accompany systemic drug administration. Also, drug loss to irrelevant pharmacokinetic compartments is decreased because the drug is administered directly to the biophase (i.e., site of action). Thus, drugs that penetrate the blood-brain barrier with great difficulty, that are rapidly metabolized, or that are available only in limited quantities can be tested for their reinforcing effects by direct cerebral microinjections.

Second, the site of action for a drug’s reinforcing effect may be localized with intracranial self-administration. Brain regions where a drug is self-administered are likely to be close to the target of drug action. This approach has several advantages over the more commonly used technique of assessing brain lesion effects on systemic drug self-administration. The latter method suffers from nonspecific lesion effects, from changes in the sensitivity of the lesioned neurons, and from the inherent limitation that lesions identify, at best, systems involved in a drug’s action but do not determine whether the drug is acting directly at that system (i.e., do not identify the area where the drug effect is initiated; see Bozarth, 1983). If appropriate control procedures are used, intracranial self-administration offers the most direct method of identifying the brain region where a drug initiates its rewarding action.

Third, if different receptor fields mediate different effects from a drug, it should be possible to minimize the influence of secondary effects on measures of drug reinforcement. For example, the peripheral sympathetic excitatory effects of psychomotor stimulants at the level of the autonomic ganglia should not be present during central microinjections of these compounds. This permits the study of stimulant reinforcement without the concomitant autonomic effects usually associated with their administration. Similarly, if the sedative or analgesic actions of opioids are due to a drug effect at brain sites that are different from those involved in their reinforcing effects, then it would be possible to study opioid reinforcement without the potentially confounding influence of sedation and analgesia (see Bozarth, 1983).

Early Studies

There were several early reports of intracranial self-administration. Cholinergic substances (Morgane, 1962; Myers, 1963) were reported to be self- administered into the hypothalamus and the septum by laboratory rats. The published reports of this work, however, gave few details of the experimental procedures. One full paper was published by Olds and Olds (1958) describing the self-administration of a monoamine oxidase inhibitor (i.e., iproniazid) into the hypothalamic area. Shortly after this promising report, a review of this work and of research involving the intracranial self-administration of other substances was published suggesting that the self-administration of these compounds resulted from nonspecific physico-chemical drug actions (J. Olds, 1962). Although this synopsis was seemingly discouraging, it laid the foundation for both the methodological requirements and some necessary control procedures for establishing the validity of intracranial self-administration experiments.

A detailed report by Olds, Yuwiler, Olds, and Yun (1964) summarized studies with 1,327 rats tested for intracranial self-administration of a wide variety of compounds. This paper emphasized basic ionic effects involved in self-administration and further illustrated the importance of controls for nonspecific physico-chemical effects following intracranial drug injections. Data showing the intracranial self-administration of cholinergic substances into the lateral hypothalamus were presented as evidence for an involvement of acetylcholine in reward processes, but even this effect was attributed to a physico-chemical action on neuronal fibers and not a direct synaptic action. During the next decade and a half, only one report was published involving intracranial self-administration. In an abstract E. Stein and J. Olds (1976) reported that morphine was intracranially self-administered at various brain sites also supporting electrical brain stimulation reward. A full report of this work was never published, but M. Olds (1979) later published a paper describing intracranial morphine self-administration into the lateral hypothalamus using the same experimental procedure.

Control Procedures

There are a several important factors to consider when studying the effects of central drug injections on behavior. These considerations demand that certain control conditions are tested to eliminate alternative explanations of the drug-taking behavior. The main factors can be conveniently divided into three broad categories: tests for behavioral, pharmacological, and anatomical specificity. Each topic addresses specific issues concerned with the interpretation of data from intracranial self-administration experiments and each suggests specific tests of the validity of such data. Although these factors have been previously described (see Bozarth, 1983), a brief summary of the basic principles involved in the assessment of intracranial self-administration studies is appropriate here.

Behavioral Specificity

The first requirement of any self-administration study is the demonstration that the animals are working for the rewarding effects of the drug injections. Both systemic and central drug injections can produce increases in locomotor activity, and such increased activity could lead to accidental lever-contacts that are mistakenly interpreted as evidence for self- administration. In particular, morphine applied to the ventral tegmental area produces increases in exploratory and locomotor activity (Joyce & Iversen, 1979) that must be considered when evaluating lever-pressing behavior. There are two primary methods that have been used to determine the relative contribution of nonspecific behavioral arousal to the lever-counts obtained during self-administration testing. The first method uses each subject as its own control, while the second method requires that two subjects be tested concurrently, one serving as the experimental subject and the other as the control subject.

The first method compares lever-press rates on two levers: one lever produces response-contingent drug injections while the other has no scheduled effect; lever-pressing on the second lever is interpreted as an indication of nonspecific lever-depressions. Although this method has been successfully used to demonstrate behavioral specificity during intracranial drug self- administration (e.g., Goeders & Smith, 1983; Monaco, Hernadez, & Hoebel, 1981), the results of this test can be variable. With behaviorally activating drug injections, the subject may depress both the active lever and depress the inactive lever during the rewarding drug injection. Thus, the rewarding effect of the drug injection would actually be paired with both levers. Because there is no penalty for pressing the inactive control lever, the discrimination learning may not be very strong. Animals may show response generalization to the inactive lever and thus increase response rates on both levers. Two-lever choice testing with intravenous heroin self-administration frequently shows excellent discrimination between active and inactive levers, but some animals show high levels of responding on the inactive lever (Bozarth, unpublished observations). When this occurs, certain subjects demonstrate behavioral specificity (i.e., preferential lever-pressing on the lever associated with drug injections), but others do not. A simple averaging of response rates on the two levers across different subjects can result in a failure to demonstrate an overall preference for the active lever. Furthermore, if a subject shows stereotypic responding on an inactive lever, the response rate on that lever can be several times the response rate on the drug-contingent lever. Averaging inflates the number of inactive lever responses obscuring the fact that other subjects may show good two-lever discrimination. The alternative procedure of using selected subjects to demonstrate two-lever choice responding can dangerously bias the experimental interpretation, although the exclusion of a small percentage of subjects stereotyping on the inactive lever is probably justified. In either case it is important that all data be presented so that the reader is free to draw his own conclusion regarding the behavioral specificity of the response.

A second method of determining if the lever-pressing is a consequence of the rewarding action of the drug is to use a yoked control procedure. With this method two animals are tested concurrently. One animal is allowed to lever press for response-contingent drug injections while the other subject is tested with an inactive lever. The lever presses of the first animal produce concurrent injections in both subjects, and the lever pressing of the second subject is simply measured as an indication of nonspecific behavioral arousal. This procedure has proven effective in studies of ventral tegmental morphine self-administration (Bozarth & Wise, 1981) as well as in studies of intravenous drug self-administration. A potential problem is that cues (e.g., activation of a cue light) accompanying the rewarding drug effect may become associated with drug reward (see Smith & Davis, this volume; Stewart & de Wit, this volume) in both subjects, and the yoked control animal may also approach these cues. This can be a problem if a cue light is illuminated directly above the lever and the control animals are repeatedly tested. Arousal and approach behavior associated with the illumination of the cue light may lead to an increase in contacts with the inactive lever. This potential problem might be eliminated by using an auditory cue associated with drug delivery. Even then, however, conditioned increases in locomotor activity may lead to increased accidental lever contacts.

Either approach to assessing the behavioral specificity of intracranial self-administration is valid. The two-lever choice test has the advantage of minimizing the number of subjects that must be tested because each subject serves as its own control. It is probably more difficult to obtain clear evidence of behavioral specificity with this method, and presentation of selected subjects demonstrating differential response rates on the two levers may bias the experimental conclusion. The yoked control procedure may provide evidence for behavioral specificity which is more easily obtainable (e.g., less likely to be influenced by response generalization producing elevated response rates on the inactive lever), but additional subjects must be tested and it may be influenced by conditioning effects. Nonetheless, reliable data from either procedure is an adequate demonstration of behavioral specificity.

Pharmacological Specificity

After the clear demonstration of behavioral specificity, the next issue of importance is the demonstration that the observed rewarding effect of the central drug injections is not the result of some nonspecific action of the compound. Microinjection of drugs into brain tissue can cause a number of nonspecific changes in the cell environment. Changes in pH, osmolarity, and regional ion balance can all occur and any of these effects might cause nonspecific activation of cells proximal to the injection site. In addition, for drugs that have rewarding effects from systemic injections, it may be important to show that the same mechanisms are involved in reward from both systemic and central drug administration.

For compounds that act at specific receptors, tests of pharmacological specificity can be conducted using receptor antagonists. In the case of opiates, these tests are particularly easy because opiate reward involves the activation of opiate receptors and because specific antagonists are available for these receptors. For example, pretreatment with an opiate antagonist such as naloxone should block the rewarding effect of central morphine if the rewarding action depends on activation of opiate receptors. Substances not having specific receptors or identified mechanisms of action (e.g., barbiturates) and substances with potent local anesthetic effects (e.g., cocaine) are more problematic. The use of active and inactive stereoisomers should reveal if the rewarding action of these compounds involves specific mechanisms or nonspecific physico-chemical actions of these substances.

When receptor antagonist or inactive stereoisomers are not available (e.g., ethanol), an alternative method of assessing pharmacological specificity must be used. If the neurochemical effect of a compound that is critically involved in its reinforcing action has been identified, then pharmacological treatments that block this neurochemical effect might be used to test the importance of this process in reward from central drug. For example, the rewarding effects of intravenous cocaine have been shown to depend on the activation of dopaminergic mechanisms, and the disruption of dopaminergic neurotransmission has been used as a test of the pharmacological specificity of reward from centrally administered cocaine. Treatment with a drug that blocks dopamine receptors attenuates intracranial cocaine self-administration into the frontal cortex (Goeders & Smith, 1983). This has been interpreted as showing that central cocaine is rewarding by the same mechanism as that involved in systemic cocaine self-administration. Caution should be exercised, however, when drawing such conclusions. Dopamine-receptor blockade has been associated with decreased locomotor activity, and the decrease in intracranial cocaine self-administration could result from a simple sedative action of the neuroleptic. Tests of active vs. inactive stereoisomers of cocaine would clearly be the preferred method of determining pharmacological specificity.

Anatomical Specificity

Anatomical specificity can be divided into two general questions: Is the rewarding drug effect due to an activation of neural mechanisms proximal to the site of injection, and how many other brain regions support the intracranial self-administration of a given compound? The first question addresses issues related to the spread of drug from the site of injection and is designed to determine if drug diffusion to a distal site of action is involved in the observed response. The second question involves anatomical mapping of the brain to determine if the observed behavior is related to a drug action in a single brain region or whether multiple brain sites can support the same response.

Physical Tests of Drug Diffusion

The diffusion of drug following intracranial injections can be assessed using methods that physically measure the quantity of drug present at different locations around the microinjection site. Radiolabeled drug can be injected and the drug spread determined with autoradiography. Serial brain sections both anterior and posterior to the injection site can be used to visualize drug dispersion, and the density of images produced can be used to assess the extent of drug diffusion. Quantitative autoradiography can provide an estimate of the actual amount of drug in different regions following microinjection. The visualization of drug dispersion with this method is time consuming because it requires serial sectioning of the brain and because adequate exposure times must be used to insure proper visualization. Furthermore, quantitative autoradiography is difficult to perform making estimates of actual drug amounts very tenuous.

Another method of physically determining the degree of drug spread following intracranial injections combines the micro-punch assay procedure of Palkovits (1973) with liquid scintillation counting. As with the autoradiographic method, the brain is sliced in serial sections. Small brain areas are then removed with a micro-punch and the amount of radioactivity present can be determined using standard liquid scintillation procedures. This method has the advantage of accurately measuring the amount of drug found at various regions around the microinjection site without introducing the complications associated with quantitative autoradiography. An excellent description of the use of this technique can be found in Myers and Hoch (1978) who used this approach to determine the dispersion kinetics of radiolabeled dopamine following intracranial injections. Although this method is relatively simple and affords an accurate assessment of drug spread, it can be very laborious if a large number of brain regions are measured.

Probably the best approach to studying physical drug spread is to perform both autoradiographic visualization and micro-punch liquid scintillation counting. Initial work could be done with autoradiography, and the results of this procedure used to direct the micro-punch assays. This way, the overall dispersion pattern will be determined and accurate quantification of drug concentrations in the target areas can also be provided. It should be noted that only autoradiography performed on a broad range of serial sections is likely to reveal drug diffusion to extremely distal sites. Such drug diffusion may occur if the microinjections encroach on cerebral vascular supplies or the cerebral ventricles. In this case drug may diffuse much further than the area likely to be included in micro-punch assays.

Functional Tests of Drug Diffusion

A different approach to determining the anatomical specificity of a response produced by intracranial injections focuses on functional tests of drug spread. Here, the physical spread of drug is considered less important, and the assessment of spread of behaviorally relevant concentrations considered the primary objective. In tests of physical drug spread, it is erroneous to conclude that the drug is acting at a distal site just because drug is present there. The behaviorally relevant concentration of drug would have to be known to make such a conclusion. Functional tests of drug spread vary the microinjection sites around the brain region found to be effective in producing the behavioral response and determine if proximal microinjections are also effective. Areas where the drug is most likely to reach by diffusion (e.g., cerebral ventricles, up the cannula shaft) are particularly important test sites (see Bozarth & Wise, 1984). If the behavior is not produced with microinjections into adjacent brain sites, then the response is likely to be produced by a drug action in the target area.

This approach has two important advantages over physical tests of drug spread. First, it is very easy to perform and does not require any facilities except those necessary to perform the intracranial self-administration experiments. The use of radiolabeled compounds requires that certain safety precautions be followed, and facilities to measure the radioactivity are required. Second, it determines if behaviorally relevant concentrations of drug are diffusing to a distal site of action. Methods measuring physical drug spread cannot determine if behaviorally relevant concentrations are being reached at distal sites but only reveal that some quantity of drug has spread to other regions.

Anatomical Mapping

One of the most important contributions that intracranial self-administration studies can make to the understanding of drug reward is the identification of brain regions where drug injections are directly reinforcing. The brain can be mapped for sites that support intracranial self-administration of a test compound using a standard protocol. If behavioral and pharmacological specificity of the intracranial self-administration has been established, this should indicate the location of receptors that initiate the rewarding actions of a given compound (see Bozarth, 1983). Tests of drug diffusion are also necessary to determine if the reinforcing effect is due to a local drug action.

The observation that a drug is not intracranially self-administered into a given brain region does not eliminate that region as a potential site for drug reward. A higher drug dosage may support intracranial self-administration, or competing responses may be produced by the central drug injections (e.g., sedation) and inhibit lever pressing. Nonetheless, mapping the brain using a standard protocol that has been shown to be effective in supporting intracranial self-administration into one or more brain regions is a first approximation to delineating the anatomical localization of reward-relevant receptors. Tests of behavioral and pharmacological specificity and tests of drug diffusion to a distal site of action must be conducted for each brain region and compound that supports intracranial self-administration.

Methodological Considerations

There are a number of methodological considerations that are important in studies of intracranial self-administration. These include the method of drug delivery and the parameters selected for studying intracranial self- administration. In this rapidly developing field, these factors have not always been given adequate attention and each laboratory appears to have developed its own standards for conducting intracranial self-administration studies.

Drug Delivery System

The topics of cannula systems and general considerations for drug microinjections have been extensively reviewed (e.g., Myers, 1972, 1974; Routtenberg, 1972) and will not be addressed here. One factor that is unique to intracranial self-administration studies and deserves special mention, however, is the method of intracranial drug delivery. Intracranial self- administration requires that low volume, response-contingent drug be delivered immediately following the lever-press response and that drug not be infused when the animal is not emitting the appropriate response. If drug is not injected immediately following the behavioral response to be reinforced or if drug is delivered at times other than immediately following the appropriate response, it is very unlikely that the subject will learn the response-reinforcement contingency. For example, a delay of reinforcement of only a few seconds is sufficient to disrupt performance in operant paradigms (see Renner, 1964; Tarpy & Sawabini, 1974). The availability of drug delivery systems that can assure the contingency of reinforcement has been perhaps the most serious limitation to the development of intracranial self-administration as a standard laboratory procedure.

Several methods are available for centrally injecting drugs into freely moving animals. The traditional approach is a direct adaptation of intravenous self-administration procedures and involves connecting the subject’s injection cannula to a microsyringe via flexible plastic tubing. A fluid commutator positioned between the microsyringe and the injection cannula permits unrestricted movement of the subject during behavioral testing. A motor-driven syringe pump advances the syringe plunger a predetermined amount following each lever press, thus displacing a known volume of drug. The presumption made by most investigators using this technique is that the amount of drug displaced from the microsyringe is centrally injected into the freely moving subject. Unfortunately, this may be an erroneous assumption.

Movement of the subject can compress and expand the plastic tubing connecting the microsyringe to the injection cannula. This can cause the injection of an undetermined quantity of drug as the subject freely moves about the test chamber. Attempts have been made to prevent the twisting and compression of the connecting tubing by using an outer cable to relieve the strain produced by the subject’s movement. The flexibility of the connecting line, however, is the reason why plastic tubing is used. Thus, making the connecting tubing sufficiently rigid to prevent stretching and compression is likely to inhibit the animal’s movement about the test chamber. Another source of potential problem with this method is the fluid swivel. If the swivel is not absolutely gas tight, leakage can occur and drug can siphon into the subject’s brain.

Various approaches have been used to increase the reliability of drug microinjections into freely moving animals. The original studies of J. Olds (e.g., 1962) used a specially developed injection technique to improve the reliability of microinjections. E. Stein and Rodd (1980) have also developed a technique that eliminates the fluid swivel which is one source of variance in drug microinjections. Both of these procedures, however, retain the flexible tubing which is likely to be the major source of variability in small volume drug delivery.

Another approach to the problem of small volume drug delivery in freely moving animals uses a technique that eliminates both the fluid swivel and flexible connecting tubing. This system uses a small gas-tight drug reservoir filled with drug and attached to an injection cannula. The entire unit is mounted directly on the subject’s head. Drug injections are accomplished by passing a small direct current between two electrodes contained in the drug reservoir. This causes the production of a specific amount of hydrogen and oxygen gas which displaces drug from the reservoir through the injection cannula. The amount of drug injected is controlled by the current intensity and duration. The animal’s injection unit is connected to the current source with light electrical leads. This eliminates the necessity of both the fluid swivel and flexible injection tubing; thus the major sources of injection volume variability are functionally circumvented while maintaining unrestricted movement of the subject during testing. (See Bozarth and Wise, 1980, for details of this procedure.) This technique has been used for studies of morphine self-administration into the ventral tegmental area (e.g., Bozarth & Wise, 1981) and for studies of cocaine self-administration into the frontal cortex (Goeders & Smith, 1983).

Testing Parameters

Intracranial self-administration studies have involved both animals previously trained to lever-press for another reinforcer and animals that were experimentally naive. Animals that are already trained to lever-press can emit high levels of responding for no reinforcer (i.e., extinction). This situation may be advantageous in initial screening for drug reinforcement from new brain sites or new compounds—the high levels of responding insure that the subject emits enough lever presses to deliver drug and may enhance the subject’s learning of the lever-press/drug-effect contingency. Also, if the drug delivery method is unreliable, high baseline rates of lever-pressing may facilitate the subject learning to respond on an intermittent reinforcement schedule. However, there is a serious limitation of using lever-trained subjects with studies involving drug effects. Animals frequently emit high levels of responding during extinction and this effect may be prolonged by certain drug treatments. Furthermore, noncontingent drug injections have been shown to reinstate lever pressing in animals only receiving saline injections (see Stewart & de Wit, this volume). Although this is likely to reveal important information regarding the nature of the drug reward, it does not establish behavioral specificity of the intracranial self-administration behavior. To satisfy this criterion, it is necessary to show that response- contingent drug delivery is maintaining the lever-pressing behavior.

Studies of direct acquisition of a lever-pressing response to receive intracranial drug provide the strongest demonstration of drug reinforcement from central injections. Although it may not always be possible to establish intracranial self-administration in experimentally naive animals, this preparation is preferred to approaches using subjects previously trained to lever press for other reinforcers.

Another important factor to consider is the concentration of the test compound injected following each lever press. In addition to the problem of nonspecific physico-chemical drug action that is usually increased with larger drug concentrations, the problem of drug spread is also exacerbated by the use of large injection doses. Drugs diffuse down their concentration gradient, and the rate and quantity of drug diffusion is determined partially by the law of mass action: The larger the concentration at the injection site the larger will be the concentration some distance from that site. Morphine is self-administered into the ventral tegmental area in picomolar unit doses (i.e., dose/microinjection), and this dose level is generally not effective in supporting intracranial morphine self-administration into other brain regions (Bozarth & Wise, 1982). Methionine enkephalin (Goeders & Smith, 1984) and cocaine (Goeders & Smith, 1983) are also self-administered into other brain sites in picomolar doses, and this suggests that this dose range may be applicable for initial tests involving the intracranial self-administration of different compounds. The use of larger injection doses may result in the diffusion of pharmacologically relevant amounts of drug to distal sites of action. For intracranial self-administration studies (as in other studies involving central drug microinjections), it is advantageous to use the lowest effective dose of a test compound. Picomolar unit doses appear to be the effective dose range for these studies. The arbitrary selection of larger injection doses is likely to invalidate any conclusions drawn regarding the anatomical specificity of the effect and may cause appreciable nonspecific effects.

The last factor to be considered here is the selection of the injection volume. In general, it might appear that the smaller the injection volume the better the anatomical resolution of intracranial microinjection studies. In fact, one study has reported a behavioral response from the microiontophoretic application of a substance (Aghajanian & Davis, 1975) suggesting that extremely small drug doses and injection volumes may be capable of producing a behavioral response. There are, however, practical considerations that also need to be taken into account. Intracranial microinjection studies require that the performance of the injection cannula be routinely checked before and after behavioral testing. For this, the volume of drug needs to be sufficient to permit easy visualization so that the injection reliability can be assured. The use of a 100 nl injection volume appears to be a reasonable trade-off between a small volume that maintains good anatomical resolution and a volume that can be visualized to determine if the microinjection system is functioning properly. Although the reliable microinjection of smaller volumes may be feasible, volumes in the low nanoliter range do not permit visual confirmation of individual drug infusions following behavioral testing. Because the single most important consideration for establishing intracranial self-administration is response contingent drug delivery, the reliability of single infusions must be assessed before and after behavioral testing. To accomplish this, it is necessary to visually confirm drug delivery following each lever press and not just the cumulative effect of numerous lever presses.


Provided that certain conditions are satisfied to insure the validity of intracranial self-administration studies, there are several advantages that can be gained by using this approach to the study of drug reinforcement. It is critical, however, that the limitations of this technique are considered in concert with the potential advantages and that the criteria for assuring the validity of these studies outlined in the previous section are fulfilled.

The first obvious advantage of using this approach is that drugs with limiting pharmacokinetic properties can be adequately tested for their rewarding effects. Compounds that do not readily cross the blood-brain barrier, that are slowly absorbed or distributed in the central nervous system, or that are rapidly metabolized or excreted are difficult to assess with systemic self-administration techniques. The direct application of these drugs into the brain should circumvent the limiting pharmacokinetic factors that preclude adequate assessment with methods such as intravenous self- administration. For example, the enkephalins and some other peptides poorly penetrate the blood-brain barrier and are rapidly inactivated. These compounds can be more effectively screened for reinforcing effects with direct microinjection into the brain.

Second, compounds that are available only in limited quantities can be screened for reinforcing effects with intracranial methods. The use of intracranial microinjection procedures should minimize the quantity of a compound required to assess its reinforcing properties. Because distribution to the critical site of action and minimal loss to irrelevant pharmacokinetic compartments result from direct drug microinjection into the brain, substantially less compound is required to produce a behavioral response. Thus, compounds that are expensive or available in limited quantities can be adequately tested if the central site of drug action is known or if it lies proximal to the cerebral ventricles.

Third, multiple actions of a drug can potentially be minimized with intracranial microinjection procedures. When a drug is injected systemically, the drug is distributed to a number of brain and peripheral sites where it may produce various effects. Many of the actions of a drug may be unrelated to its rewarding action and some of these side-effects may even obscure the important rewarding effects of a drug. For example, when opiates are intravenously self- administered, drug is distributed to a number of brain sites where several different actions are produced. A periventricular gray action of opiates produces analgesia (Sharpe, Garnett, & Cicero, 1974; Yaksh, Yeung, & Rudy, 1976) and if the dose is sufficient, physiological dependence (Bozarth & Wise, 1984) and sedation (Pert, DeWald, Liao, & Sivit, 1979) may occur. Changes in thermoregulation can be produced by a drug action in the preoptic area of the lateral hypothalamus (Lotti, Lomax, & George, 1965; Teasdale, Bozarth, & Stewart, 1981), and effects on prolactin secretion can accompany an opiate action in the raphe nucleus (Johnson, 1982). The rewarding and locomotor stimulating actions of opiates appear to be derived from an action in the ventral tegmental area which does not involve these other opiate effects (see Bozarth, 1983). Thus, the compound opiate effects produced during systemic drug administration are related to the drug’s actions on multiple brain sites and most of these effects are not related to the rewarding action of the compound. Central injections of opiates can isolate the rewarding action from analgesia, physiological dependence, and other opiate effects that may be irrelevant to opiate reinforcement (see Bozarth, 1983).

Fourth, the independent contributions of different reinforcement systems simultaneously activated by systemic drug delivery can be assessed with this technique. Opiates, for example, appear to involve a strong positive reinforcing action in the ventral tegmental area (Bozarth & Wise, 1981, 1982; van Ree & De Wied, 1980). This rewarding effect is independent of physiological dependence mechanisms as revealed by the fact that repeated morphine injections into this region fail to produce physiological dependence (Bozarth & Wise, 1984). It is possible, however, that physiological dependence contributes to the net reinforcing impact of systemically delivered opiates in subjects that have become physically dependent. The relief of withdrawal distress may provide negative reinforcement and may be capable of maintaining lever-pressing behavior. By limiting the distribution of morphine to the brain region where positive reinforcement is initiated (i.e., ventral tegmental area) or the area where relief of withdrawal discomfort is likely to be produced (i.e., periventricular gray region), the contribution of positive and negative reinforcement processes from these two brain systems can be independently evaluated. Similarly, if multiple brain systems are involved in the positive reinforcing actions of a compound (e.g., frontal cortex and nucleus accumbens in the case of cocaine), then the relative importance of each system can be evaluated with the intracranial self-administration paradigm. If several brain sites are capable of initiating the reinforcing effect of a drug, then brain lesion challenges of a single site during intravenous self-administration would be unlikely to reveal the importance of that site in drug reward because the unlesioned brain system may be capable of maintaining the rewarding action of the test compound. Intracranial self-administration permits the systematic determination of the ability of various brain regions to initiate the rewarding actions of a compound and can thus identify brain systems involved in drug reinforcement even if multiple systems are capable of maintaining the behavior.

There are several serious limitations of intracranial self-administration procedures. The most obvious is that this technique is methodologically difficult requiring the precise delivery of nanoliter volumes of drug in freely moving animals. Improvements in microinjection technology can eliminate this problem, but caution needs to be observed when studying the effects of repeated drug microinjections because of the potential problem of tissue trauma. The repeated microinjection of substances in the brain can radically alter the cell environment, and changes in the responsiveness to drug can occur because of physiological changes produced by the injection procedure. These problems, however, can be minimized by the selection of an adequate microinjection procedure and by the use of low volume and low infusion rate injections that are not traumatic to neural tissue.

A limitation of intracranial self-administration studies that is more difficult to circumvent involves the nature of brain reinforcement systems. It is possible that reinforcement from a given drug might involve the concurrent activation of several brain sites that cannot be duplicated by drug injection into a single site. Thus, conventional microinjection studies which examine the effect of a drug injection into one brain site at a time would not reveal a reinforcing action for a compound that requires the concurrent activation of several neural systems. Nonetheless, a number of compounds have been shown to be reinforcing when microinjected into discrete brain regions thus suggesting that reinforcement from at least some substances can result from a drug action initiated at a single brain region (see Table 1).

Table 1
Compounds Intracranially Self-Administered
Compound Brain Site Investigator(s)
amphetamine FCX Phillips et al., 1981
  NAS Monaco et al., 1981
    Hoebel et al., 1983
calcium chelators LHA J. Olds et al., 1964
cholinergic agent(?) PFR Morgane, 1962
  Spt Myers, 1963
cocaine  FCX Goeders & Smith, 1983
fentanyl VTA van Ree & De Wied, 1980
iproniazid LHA J. Olds & M. Olds, 1958
methionine enkephalin NAS Goeders & Smith, 1984
d-ala2-met-enkephalinamide LHA M. Olds & Williams, 1980
morphine LHA + others E. Stein & J. Olds, 1977
  LHA M. Olds, 1979
  NAS M. Olds, 1982
  VTA Bozarth & Wise, 1981, 1982
Abbreviations: FCX, frontal cortex; LHA, lateral hypothalamic area; NAS, nucleus accumbens; PFR, perifornical region; Spt, septal region; VTA, ventral tegmental area. One limitation of the intracranial self-administration paradigm cannot be circumvented and is inherent in the very nature of this approach to studying drug reinforcement. Intracranial self-administration demonstrates that a drug action in a specific brain region is sufficient for reinforcement. That is, activation of that region by the compound can directly reinforce behavior. The fact that a compound is self-administered into a given brain region does not, however, suggest that activation of that brain region is necessary for reinforcement from systemic application of the test compound (see Bozarth, 1983). It could be the case that several brain systems are involved in the reinforcing action of a single compound and that activation of any one of these systems is sufficient to reinforce behavior. Only studies assessing the effects of blocking drug action in a given brain region during tests of the reinforcing effects of systemically administered drug (e.g., intravenous self-administration) can determine if a drug action in a given brain region is necessary for drug reinforcement.

Intracranial self-administration procedures can provide an important advance in the study of drug reinforcement mechanisms. This paradigm is uniquely suited for determining the anatomical localization of brain sites responsible for the initiation of drug reward and can provide important corroboration of findings revealed with other methods. Intracranial self- administration is a relatively new and unexplored technique, although the first studies using this approach were reported thirty years ago. Because the standards for evaluating these experiments have not been clearly established in the literature, it is especially important to be cautious when interpreting the results of these studies. The criteria of behavioral, pharmacological, and anatomical specificity provide minimal requirements for demonstrating the validity of these studies. In addition to the acceptance of uniform standards for conducting these experiments, further improvements in microinjection technology will undoubtedly permit the more widespread application of this important procedure for studying drug reinforcement.


Preparation of this manuscript was supported by grants from the Natural Science and Engineering Research Council of Canada (NSERC) and by the National Institute on Drug Abuse (U.S.A.). The author is a University Research Fellow sponsored by NSERC.


Aghajanian, G.K., & Davis, M. (1975). A method of direct chemical brain stimulation in behavioral studies using microiontophoresis. Pharmacology Biochemistry & Behavior, 3, 127-131.

Bozarth, M.A. (1983). Opiate reward mechanisms mapped by intracranial self- administration. In J.E. Smith and J.D. Lane (Eds.), Neurobiology of opiate reward processes (pp. 331-359). Amsterdam: Elsevier/North Holland Biomedical Press.

Bozarth, M.A., & Wise, R.A. (1980). Electrolytic microinfusion transducer system: An alternative method of intracranial drug application. Journal of Neuroscience Methods, 2, 273-275.

Bozarth, M.A., & Wise, R.A. (1981). Intracranial self-administration of morphine into the ventral tegmental area in rats. Life Sciences, 28, 551-555.

Bozarth, M.A., & Wise, R.A. (1982). Localization of the reward-relevant opiate receptors. In L.S. Harris (Ed.), Problems of drug dependence, 1981 (National Institute on Drug Abuse Research Monograph 41, pp. 158-164). Washington, DC: U.S. Government Printing Office.

Bozarth, M.A., & Wise, R.A. (1984). Anatomically distinct opiate receptor fields mediate reward and physical dependence. Science, 244, 516-517.

Goeders, N.E., & Smith, J.E. (1983). Cortical dopaminergic involvement in cocaine reinforcement. Science, 221, 773-775.

Goeders, N.E., Lane, J.D., & Smith, J.E. (1984). Self-administration of methionine enkephalin into the nucleus accumbens. Pharmacology Biochemistry & Behavior, 20, 451-455.

Hoebel, B.G., Monaco, A.P., Hernandez, L., Aulisi, E.F., Stanley, B.G., & Lenard, L. (1983). Self-injection of amphetamine directly into the brain. Psychopharmacology, 81, 158-163.

Johnson, J.H. (1982). Release of prolactin in response to microinjection of morphine into mesencephalic dorsal raphe nucleus. Neuroendocrinology, 35, 169-172.

Joyce, E.M., & Iversen, S.D. (1979). The effect of morphine applied locally to mesencephalic dopamine cell bodies on spontaneous motor activity in the rat. Neuroscience Letters, 14, 207-212.

Lotti, V.J., Lomax, P., & George, R. (1965). Temperature responses in the rat following intracerebral microinjections of morphine. Journal of Pharmacology and Experimental Therapeutics, 150, 135-139.

Monaco, A.P., Hernandez, L., & Hoebel, B.G. (1981). Nucleus accumbens: Site of amphetamine self-injection: Comparison with the lateral ventricle. In R.B. Chronister and J.F. DeFrance (Eds.), Neurobiology of the nucleus accumbens (pp. 338-343). Brunswick, ME: Haer Institute.

Morgane, P.J. (1962). Reinforcing effects of self-injected cholinergic agents into hypothalamic ‘drinking’ areas in rats. Federation Proceedings, 21, 352.

Myers, R.D. (1963). An intracranial chemical stimulation system for chronic or self-infusion. Journal of Applied Physiology, 18, 221-223.

Myers, R.D. (1972). Methods for chemical stimulation of the brain. In R.D. Myers (ed.), Methods in psychobiology (Vol. 1, pp. 247-280). New York: Academic Press.

Myers, R.D. (1974). Handbook of drug and chemical stimulation of the brain. New York: Van Nostrand Reinhold.

Myers, R.D., & Hoch, D.B. (1978). C-dopamine microinjected into the brain-stem of the rat: Dispersion kinetics, site content and functional dose. Brain Research Bulletin, 3, 601-609.

Olds, J. (1962). Hypothalamic substrates of reward. Physiological Reviews, 42, 554-604.

Olds, J., & Olds, M.E. (1958). Positive reinforcement produced by stimulating hypothalamus with iproniazid and other compounds. Science, 127, 1175-1176.

Olds, J., Yuwiler, A., Olds, M.E., & Yun, C. (1964). Neurohumors in hypothalamic substrates of reward. American Journal of Physiology, 207, 242-254.

Olds, M.E. (1979). Hypothalamic substrate for the positive reinforcing properties of morphine in the rat. Brain Research, 168, 351-360.

Olds, M.E. (1982). Reinforcing effects of morphine in the nucleus accumbens. Brain Research, 237, 429-440.

Olds, M.E., & Williams, K.N. (1980). Self-administration of d-ala-met-enkephalinamide at hypothalamic self-stimulation sites. Brain Research, 194, 155-170.

Palkovits, M. (1973). Isolated removal of hypothalamic or other brain nuclei of the rat. Brain Research, 59, 449-450.

Pert, A., DeWald, L. A., Liao, H., & Sivit, C. (1979). Effects of opiates and opioid peptides on motor behaviors: Sites and mechanisms of action. In E. Usdin, W. Bunney, Jr., and N. S. Kline (Eds.), Endorphins in mental health research (pp. 45-61). London: MacMillan.

Phillips, A. G., Mora, F., & Rolls, E. T. (1981). Intracerebral self- administration of amphetamine by rhesus monkeys. Neuroscience Letters, 24, 81-86.

vanRee, J. M., & De Wied, D. (1980). Involvement of neurohypophyseal peptides in drug-mediated adaptive responses. Pharmacology Biochemistry & Behavior, 13(Suppl. 1), 257-263.

Renner, K. E. (1964). Delay of reinforcement: A historical review. Psychological Bulletin, 61, 341-361.

Routtenberg, A. (1972). Intracranial chemical injection and behavior: A critical review. Behavioral Biology, 7, 601-642.

Sharpe, L. G., Garnett, J. E., & Cicero, T. J. (1974). Analgesia and hyperactivity produced by intracranial microinjections of morphine into the periaqueductal gray matter of the rat. Behavioral Biology, 11, 303- 313.

Stein, E. A., & Olds, J. (1977). Direct intracerebral self-administration of opiates in the rat. Society for Neuroscience Abstracts, 3, 302.

Stein, E. A., & Rodd, D. (1980). A new injection technique for intracerebral drug administration in behaving animals. Pharmacology Biochemistry Behavior, 12, 815-817.

Tarpy, R. M., & Sawabini, F. L. (1974). Reinforcement delay: A selective review of the last decade. Psychological Bulletin, 81, 984-997.

Teasdale, J. A. P., Bozarth, M. A., & Stewart, J. (1981). Body temperature responses to microinjections of morphine in brain sites containing opiate receptors. Society for Neuroscience Abstracts, 7, 799.

Yaksh, T. L., Yeung, J. C., & Rudy, T. A. (1976). Systematic examination in the rat of brain sites sensitive to the direct application of morphine: Observation of differential effects within the periaqueductal gray. Brain Research, 114, 83-103.

©1987 Springer-Verlag (printed version)
©2000-2009 Addiction Science Network (web-enhanced version)

ASNet Home

ASNet Profile

Illicit Drug Index

Virtual Lab Tour

Research Reports

Drug Classification

A Primer on Drug Addiction

Experimental Methods

Treatment Resources

Biological Basis of Addiction

Before Prohibition: Early Psychoactive Medicines

Links to Other Websites

Click here to enter the Addiction Science Network Discussion Forum

Brain Reward System©1999-2009 Addiction Science Network
This page was last revised 06 April 2009 13:06 EDT.
Send comments to:
Report technical problems to: