Saturday, June 9, 2012

Towards Designing an Ideal Intervention For Substance Abuse: Key Reference Points from Some of the Studies Discussed Thus Far

While much work still remains to be done in such areas as delineating the specific neural circuitry involved in drug addiction, several factors involved in addiction- and the implications they present for designing an optimal treatment program- have become clear.

For one thing, the Opponent Process Theory of Motivation, first discovered by Solomon and Corbit (1974) has made clear that addicts self-administer drugs not because they want to attain any pleasurable high that might come as a result of doing so, but rather so as to avoid the devastating effects of the B-process that are sure to strike them down should they cease self-administering drugs. That this is the case is also evident from a study by a study by Miller, Klahr, Gold, Cocores, & Sweeney (as cited by Budney, Higgins, Delaney, Kent & Bickel, 1991 ), in which the researchers found that many of the reasons the addicts in that study cited for self-administering a second drug in fact had to do with their wanting to avoid the b-process effects of a first drug. From this, it becomes evident that therapies such as Cognitive Behavioral Therapy, which largely focus upon helping a patient to gain insight into the causal factors, (such as maladaptive thought patterns and the like) responsible for their unacceptable behavior are unlikely to be particularly helpful. A similar comment might be made in regard to therapies that aim to facilitate the overcoming of drug addiction by increasing a patient’s motivation to do so; rather, it would appear that what is needed is a more direct means of behavior modification.

The community reinforcement approach advocated for by Higgens, Delaney, Budney, Hughes, Foeng, and Fenwick (1991) appears to be a step in this correct direction, at least in terms of helping individuals to maintain their independence from substance abuse addictions (once this has been acquired) and preventing relapse. As described  in a previous blog post, this means of therapy “reinforces” correct behavior (in the form of abstinence from drugs) by providing opportunities for previously-addicted individuals to become increasingly integrated into their communities as functioning and contributing members,  thus eliminating both the need for expensive and ultimately ineffective external motivating forces and, perhaps, (although more indirectly) the need for individuals to take drugs in the first place (to the extent that the conventional wisdom that individuals first begin taking drugs in order to counteract the effects of what they perceive as unbearable stress or highly aversive life circumstances is true).

As an initial means of helping individuals to gain independence from substance abuse addictions, the conditioned taste aversion method employed in the study conducted by Howard (2001) appears to be both effective in terms of achieving treatment results, and cost-effective, with results becoming evident after only a single trial. As a taste-aversion method, it also carries the benefit of the fact that learning can still occur even if nausea is brought on several hours after the injestion of alcohol, as well as the fact that, since it relies upon associations [such as that between taste and gastrointestinal discomfort, which, as Garcia and Koelling (1966) discovered in their experiment concerning CS-US belongingness, the human brain is particularly inclined to make], making it a particularly good candidate. While this experiment by Howard (2001) is, unfortunately, lacking in control groups and needs to be replicated before definitive conclusions can be drawn, preliminary evidence- such as the widespread acceptance of the alcohol-cessation aid antabuse, which functions according to similar principles as Howard's (2001) taste aversion conditioning treatment, points to its effectiveness. Furthermore, while the external validity of such a treatment might be limited (it is doubtful that anyone in the outside world could be reliably be expected to artifically induce gastrointestinal illness in a recovering alcoholic every time such an individual has consumed a dangerously high amount of alcohol), perhaps its effects, once attained through treatment, can be maintained by such intrinsic-motivation-enhancing means as the community reinforcement approach advocated for by Higgins et al. (1991).

The study by Caddy and Lovibond (1976) adds to this emerging picture of the "ideal" (insofar as has become clear from these previously-discussed studies) treatment for alcohol addiction the principle of the employment, during treatment, of a variable-ratio schedule of reinforcement for punishing the consumption of alcohol above a previously-prescribed limit. The employment of such a variable-ratio schedule is important as it both mimicks the schedule of reinforcement according to which illicit behaviors are most likely to be punished in real life, and, [as is further discussed in a previous blog post discussing the work of Conklin and Tiffany (2002)], and as I also learned about over the time spent in my Fundamentals of Learning course, it is a schedule or reinforcement that results in the establishment of behavior that is robust even in the face of extinction conditions.

Finally, the study by Robinson and Berridge (2003), which led to the key finding that, in the words of Robinson and Berridge (2003), "Once the b-process is strengthened, even a small dose of the drug can instate it and thereby trigger withdrawal again. Conversely, prolonged abstinence from the drug would decay the b-process, and the ability to reactivate it would return back to normal. Once the b-process returns back to normal, the person would no longer be addicted," (Robinson and Berridge, 2003, p. 28). First, this finding collaborates the work of other researchers in confirming that a key feature of any substance abuse program is that of ensuring that the recovering addict goes through an adequate period of time with no access whatsoever to the drug to which they were previously addicted. As, for instance, the work of  Ehrman, Robbins, Childress, and O'Brien (1992) demonstrated, exposure to even drug-related conditioned stimuli alone can inspire strong cravings and withdrawal symptoms among former drug addicts. Furthermore, the "decay of the b-process" (Robinson and Berridge, 2003, p. 28) which the authors make reference to also goes some distance in explaining the phenomena in which some addicts, having just completed a rehabilitiation program, suddenly experience what they perceive as an unbearable craving for even a relatively small dosage  of their drug of choice. Unfortunately, for many of these individuals, this final self-administration of the drug of choice results in an overdose (due to the decay of the previously-built-up tolearance). Unfortunately (at least in my very limited experience of reading fiction literature in which such deaths occasionally occur), these seem to often be labeled simply "overdoses," without any additional explanation being added to differentiate them from what one might intuitively understand an "overdose" to be. If this is also the case in the real world, then perhaps greater education for both recovering  substance abusers and the general public in regards to the decay of the b-process and the heightened risk of death it presents for addicts just coming out of rehabilitation facilities and who might be tempted to injest their drug of choice "just one last time," might be useful. While of course (as convergent evidence from many of the studies discussed in previous blog posts suggests) the ideal for any recovering substance abuser is total abstinence from the drug of choice, given that the average relapse rate for substance abusers is still currently far higher than zero, perhaps some basic education in regards to the b-process and the breakdown of previously built-up tolerance might be instrumental in saving the lives of some recovering substance abuse addicts.

 Another key finding which the study by Robinson and Berridge (2003) resulted in was that a sudden cessation in the self-administration of a particular drug can lead to a precipitious drop in levels of the brain neurotransmitters dopamine and serotonin. which, combined with the overactivation of the hypothalamic-pituitary-adrenocortical axis which can also occur at the same time can often throw newly-recovering addicts into such a state of physiological stress that they become particularly vulnerable to relapse. Indeed, Robinson and Berridge (2003) present such stress effects as being strong enough to be comparable in strength to a former addict suddenly being exposed to a small amount of the drug to which they were previously addicted  (which, as the authors mention, would go on to activate the b-process and result in strong withdrawal symptoms and drug cravings). Given that the effects of stress are so potent in stimulating relapse, perhaps treatment programs of the future could incorporate some form of "graduated re-entry"into the outside world for patients newly leaving inpatient treatment facilities, perhaps by insuring that patients are enrolled in some sort of program along the lines of the community reinforcement approach programs discussed by by Higgins et al. (1991) before they ever leave the treatment facility.

Budney, A. J., Higgins, S. T., Delaney, D. D., Kent, L., & Bickel, W. K. (1991). Contingent
reinforcement of abstinence with individuals abusing cocaine and marijuana. Journal of
Applied Behavior Analysis, 24(4), 657-665. doi: 10.1901/jaba.1991.24-657. Retrieved

Caddy, G. R., & Lovibond, S. H. (1976). Self-regulation and discriminated aversive conditioning in the modification of alcoholic's drinking behavior.Behavior Therapy, 7(2), 223-230. doi: 10.1016/S0005-7894(76)80279-1

Conklin, C. A., & Tiffany, S. T. (2002). Applying extinction research and theory to cue-exposure addiction treatments. Addiction, 97(2), 155-167. doi: 10.1046/j.1360-0443.2002.00014.x

Ehrman, R. N., Robbins, S. J., Childress, A. R., & O'Brien, C. P. (1992). Conditioned responses to cocaine-related stimuli in cocaine abuse patients.Psychopharmacology, 107(4), 523-529. doi: 10.1007/BF02245266

Garcia, J., & Keolling, R. A. (1966). Relation of cue to consequence in avoidance learning. Psychonomic .Sci., 4, 123-124. Retrieved from

Higgins, S. T., Delaney, D. D., Budney, A. J., Bickel, W. K., Hughes, J. R., Foerg, F., & Fenwick, J.
W. (1991). A behavioral approach to achieving initial cocaine abstinence. The American
Journal of Psychiatry, 148(9), 1218-1224. Retrieved from

Howard, M. O. (2001). Pharmacological aversion treatment of alcohol dependence.i. production and prediction of conditioned alcohol aversion. The American Journal of Drug and Alcohol Abuse,27(3), 561-585. doi: 10.1081/ADA-100104519

Robinson, T. E., & Berridge, K. C. (2003). Addiction. Annual Review of Psychology, 54, 25-53. doi: 10.1146/annurev.psych.54.101601.145237

Solomon, R. D., & Corbit, J. D. (1974). An opponent-process theory of motivation: I. temporal dynamics of affect. . Psychological Review, 81(2), 119-145. doi: 10.1037/h0036128

Saturday, June 2, 2012

       In their article "Applying Extinction Research and Theory to Cue-Exposure Addiction Treatments," Conklin and Tiffany discuss the effectiveness of the cue-exposure extinction treatments currently in use for disorders such as substance abuse and other addictions, and suggest directions for improvement, based on new research performed with animals.
Conklin and Tiffany (2002) begin by stating that it has been known for a while now that cue exposure is an effective means of treatmeant for individuals addicted to drugs (Hammersley, 1992).  As evidence for this, they present the fact that "... drug-use and relapse are often strongly cue and context-specific" (Conkin and Tiffany, 2002, p. 155). This theory is compatible with commentary offered by Domjan (2009), in which he states that "It has been known for a long time that the administration of a drug constitutes a conditioning trial which cues related to drug administration are paired with the pharmacological effects of the drug," (Domjan, 2009, p. 115). It is also consistent with evidence obtained from a study by Siegel (2000), in which he found that an initial self-administration of a smaller amount of a drug (which, among drug addicts, is commonly followed-up by the self-administration of a larger amount of the same drug, leading to the initial, smaller amount of the drug becoming a conditioned stimulus predicting an impending administration of a larger amount of the same drug). When these lines of evidence are taken together, it becomes clear that, if the power of contextual cues related to drug self-administrations to elicit cravings could somehow be extinguished, individuals formerly addicted to drugs might have an easier time remaining abstinent from addictive substances. The present study by Conklin and Tiffany (2002) explores how this might be accomplished.

       Conklin and Tiffany (2002) go on to describe how the process of cue-exposure treatment for drug addiction currently proceeds. According to them, it involves repeated exposure to contextual stimuli associated with drug use, with the crucial change that, unlike in the addict's previous episodes of drug self-administration, in these instances, presentations of such cues would not be reinforced with the administration of the drugs themselves. The hope is that, eventually, such repeated exposure to the cues and the unexpected absence of the drug itself will lead to extinction of the drug cravings normally triggered by drug cues.
       Evidence that drug-related cues do indeed trigger drug-related cravings in the manner described by Conklin and Tiffany (2002) has been obtained from a study performed by Ehrman, Robbins, Childress, and O'Brien (1992). That study involved two groups of men: one group who had a history of using cocaine (but not heroin), and a control group of men who had no history of using either cocaine or heroin. The study involved exposing the participants to several conditions, including that of exposing them to cues related to cocaine, through such means as having them listen to audiotapes and watch videos of other individuals involved in cocaine-related actions, as well as having them pretent to use cocaine by acting out the motions involved in doing so. The participants were then exposed to another test condition in which they performed the same actions and were exposed to the same cues, although this time as related to the use of heroin (a drug to which they had never been addicted, and, thus, one to which they should show a greatly attenuated reaction, in comparison with their reaction to cocaine). Finally, the participants were exposed, in the same manner, to a set of control stimuli unrelated to the use of any drug. The results of the study showed that only the participants with a history of cocaine use showed a strong increase in heart-rate as well as a high degree of self-reported cravings for cocaine and symptoms of cocaine withdrawal. The fact that these symptoms were not evident among the same participants when they were exposed to the other two sets of stimuli (those related to heroin, which they did not have a history of using, and those related to neutral, non-drug-related control stimuli), indicates that drug-related cues trigger strong drug cravings for a drug to which those cues are related, and that this occurs only among individuals with a previous history of having used that particular drug. This serves as one line of evidence justifying Conklin and Tiffany's (2002) emphasis of extinguishing drug cravings among drug addicts by extinguishing the power of drug-related cues to elicit drug-related cravings.

       Conklin and Tiffany (2002) present the rationale for the use of cue-exposure treatments for drug addiction as being rooted in a classical conditioning model, in which "the drug is the unconditioned stimulus (US) and the drug effects are the unconditioned responses (UR). The conditions under which the drug is used become conditioned stimuli (CS) that evoke conditioned responses (CR) that moderate or mediate drug seeking and drug consumption," (Conklin and Tiffany, 2002, p. 156). Such a conceptualization is consistent with the above-cited studies by Siegel (2000) and Ehrman et al. (1992).
 Interestingly, research by O'Brien, Childress, McLellan, and Ehrman (1990) has also found that the procedures designed to extinguish the potential of drug-related stimuli to elicit drug-related cravings can indeed be successful in humans; in that study, thirty participants who had formerly been dependent on cocaine but had now been abstinent for some time were initially exposed to two conditions: one in which they were exposed to cues related to the self-administration of cocaine, and another condition in which they were exposed to neutral stimuli. The results of these baseline measures indicated that the participants showed a heightened sensitivity and responsiveness to cues related to the self-administration of cocaine, as compared to the neutral stimuli. In response to exposure to cues related to cocaine self-administration, the participants experienced heightened physiological arousal, along with increases in self-reported symptoms of cocaine cravings, cocaine-related "highs," and withdrawal symptoms. Significantly, these reactions were not present during their exposure to the neutral control stimuli. During the treatment phase of the experiment, the participants were exposed to a cue-exposure addiction treatment in which they were repeatedly exposed to cues related to the self-administration of cocaine, without actually being exposed to the drug itself. Such treatment was performed for a total of fifteen one-hour sessions. Crucially, the participants reported a complete extinction cocaine-related cravings by the end of the fifteenth session of cue-exposure treatment, indicating that the cue-exposure treatment had been successful in eliminating the capacity of the cocaine-related stimuli to stimulate drug-related cravings in these participants! Furthermore, the experiences of "highs" and cocaine-related "withdrawal symptoms" which the participants had initially reported were eliminated by the end of the sixth week of treatment- a fact serving as further evidence of the efficacy of this treatment. Furthermore, the results of that study indicated that participants in the experimental group (the ones who received this cue-exposure treatment) showed a greater ability to maintain the results obtained over the course of treatment than did the participants in the control group, who had been exposed to standard treatment for cocaine addiction alone (i.e.,without the addition of the cue-exposure treatment), (O'Brien et al., 1990). These results indicate that the cue-exposure treatment therapy discussed by Conlin and Tiffany (2002) is indeed promising.

Interestingly, Conklin and Tiffany (2002) also note that the methods by which researchers have orchestrated cue-exposure therapy for drug addiction have largely been consistent across different types of addictive substances. In their words, "Typically, addicts are exposed to personally-relevant drug cues either in vivo (i.e., handling drug paraphernalia) or imaginal (i.e., imagining being in a situation typical of past drug use)," (Conklin and Tiffany, 2002).
       Given this, it might be interesting to conduct an investigation in regards to which of these two methods of exposure is more effective. For instance, if the objective, physical exposure to drug-related cues turns out to be more effective, then cue-exposure treatment should, in the future, focus mainly on this method of exposure; if, by contrast, the imaginal treatment turns out to be equally or perhaps even more effective, then this might be an indication that the roots of drug addiction are more cognitive in nature, and that cue-exposure treatment should include more cognitive and perhaps even more motivational components in the future.

In their review of past studies of cue-exposure treatment for drug addiction, Conklin and Tiffany include fifteen studies involving a variety of addictive substances, including "opiates (N= 6), nicoteine (N= 5), alcohol (N= 5) and cocaine (N=1), " (Conklin and Tiffany, 2002, p. 157). They also note that one study included in their review involved both participants addicted to cocaine and participants addicted to opiates. In regards to their selection criteria, Conklin and Tiffany (2002), "... any treatment involving exposing addicts to cues associated with past drug use in an attempt to extinguish learned responding to those cues," (Conklin and Tiffany, 2002, p. 156), although they did not include "... studies [that] involved exposure to drug cues for the purpose of replacing a learned response with an alternative response (e.g., covert sensitization)," (Conklin and Tiffany, 2002, p. 157). Furthermore, they note that while approximately half of the studies were conducted with individuals maintaining abstience from the addictive substances during the course of the study, many of the studies involved participants who were still involved in ingesting the drug of abuse during the course of the study (the specific reasons for this varied widely across studies). The most prevalent mode of presentation of drug cues over the course of the studies was "in vivo," (Conklin and Tiffany, 2002, p. 157), a mode of presentation which was used in fifteen of the eighteen studies. Conklin and Tiffany (2002) operationally defined this mode of drug cue presentation as "... including handling one's own or simulated drug paraphernalia and drug itself, ingestion of actual drug (e.g. priming doses of alchol) or simulated drug (e.g. injecting saline), preparing drugs for use (e.g. lighting a cigarette, 'cooking up' and 'tying off'), and outside exposure in an environment associated with past drug use," (Conklin and Tiffany, 2002, p. 157). Other modes of presentation (which were often presented in combination with that mode), included the presentation of photographic cues (i.e., pictures of addicts actively engaged in abusing drugs), video cues (i.e., videos of addicts onvolved in obtaining or ingesting drugs), audio cues (i.e., those involving recordings of individuals discussing their drug-taking experiences), as well as imaginal cues (which involved asking participants to visualize a drug-taking experience). While the number of cue-exposure sessions in the studies reviewed by Conklin and Tiffany (2002) varied, as did the length of the treatment sessions and the number of cues a participant was exposed to over the course of any single session, the studies were consistent in that they used one of three possible methods to determine the length of time for which participants in a study were exposed to any partilcular drug-related cue. Specifically, the three possible methods for determining how long a participant was to be exposed to a particular drug-related cue included "(1) a specific amount of time was predetermined... (2) a specific acttion was required and when it was completed exposure to that cue ended... and (3) the treatment was terminated when the participant's self-reported craving/ urge level dropped to half the peak intensity experienced during exposure to the target cue," (Conklin and Tiffany, 2002, p. 157). Furthermore, "Treatment sessions either occurred for a set number of days... or were distributed across a specific number of days," (Conklin and Tiffany, 2002, p. 157).

Conklin and Tiffany (2002) evaluated the effectiveness of the cue-exposure treatments among the studies they reviewed by "...applying meta-analytical techniques to the abstinence or drug-use reduction results from each treatment outcome study," (Conklin and Tiffany, 2002, p. 157). All of the studies included in their meta-analysis met the following two selection criteria: "(1) the study included a control or comparison treatment group and (2) a post-treatment follow-up, during which abstinence or drug use was measured, was reported for each group," (Conklin and Tiffany, 2002, p. 159). Conklin and Tiffany (2002) then used "DSTAT, a computer program for meta-analytical reviews of research literature," (Conklin and Tiffany, 2002, p. 159), to compute the effect sizes for each of the studies represented in their review. As Conklin and Tiffany (2002) went on to explain, "Effect sizes showing greater abstinence or less drug use for the cue-exposure treatment group were assigned a positive value, while negative values denoted less abstinence or greater drug use for the cue-exposure group," (Conklin and Tiffany, 2002, p. 159). These effect sizes were then corrected for bias and tested for homogeneity. Results indicated that "The overall effect-size of cue-exposure treatment (d= 0.0868; 95% confidence interval - 0.11 ± 0.28) was not significant. The Q statistic reached significance (Q(9)= 16.078; p= 0.0413), indicating that effect sizes were not consistent across studies," (Conklin and Tiffany, 2002, p. 159).

In light of these results indicating that cue-exposure treatment for preventing relapse among drug addicts is, at the present time, less than efficacious, Conklin and Tiffany (2002) go on to present suggestions for its improvement. They justify such efforts with their commentary that "... there is a pervasive belief that if the optimal parameters for cue-exposure could be discovered (e.g. the right cues are chosen, the best number of sessions are conducted, it is combined with the right psychotherapy) addiction treatment might have a new 'gold standard' for treatment efficacy," (Conklin and Tiffany, 2002, p. 159), and they present recently-obtained results from animal studies involving extinction of learned behavior as an indication of the best means to go about doing so.

Conklin and Tiffany (2002) go on to descrbe how, originally, extinction was viewed as a weakening or undoing of a CS-US association first learned during acquisition. This is the way extinction had been conceptualized by, for instance, by Rescorla and Wagner (1972), as cited in Conklin and Tiffany (2002). Conklin and Tiffany (2002) go on to explain that, currently, extinction is viewed not as an unlearning of a previously-conditioned association, but rather with the idea that, "...during extinction, CS-US learning remains intact, but new associations develop to the original CS," (Conklin and Tiffany, 2002, p. 159). As Conklin and Tiffany (2002) go on to explain, "From this perspective, the effectiveness of extinction, and similarly the efficacy of cue exposure, is determined by the probability that exposure to an extinguished drug cue post treatment will evoke behavior learned during extinction (i.e., abstinence) rather than that learned during original conditioning (i.e., drug use), (Conklin and Tiffany 2002, p. 159).
This thesis is in line with Domjan's (2009) commentary that, rather than being simply the unlearning of an originally-acquired association, extinction involves the learning of a new "inhibitory S-R association," (Domjan, 2009, p. 300). According to Domjan, the inhibitory nature of this newly-formed association arises from the "frustrative effects of the unexpected absense of reward," (Domjan, 2009, p. 300).
An example of how extinction conditions [which, as Domjan (2009) explains, can involve either a lack of presentation of a US after the presentation of the CS normally predicting that US, or a failure to reward an instrumnetal response with the reward that is normally contingent upon the organism making that response] can result in frustration, and, ultimately, in the inhibition of behavior is presented in a study performed by (Neuringer, Kornell, and Olufs, 2001). That study involved two groups of rats, divided into an experimental and control group. During the acquisition phase, all of the rats were rewarded for each series of three responses they made using the levers and other equipment provided. However, the rats in the experimental group were reinforced only if they varied the order of the three responses they made, whereas the rats in the yoked control group were rewarded for making three responses, regardless of the specific order in which they made them. After the rats had developed a well-established pattern of responding, the second phase of the experiment began. In this phase, all of the rats were shifted to extinction conditions, in which they were not reinforced with food no matter what pattern of lever-presses they employed. The results indicated the expected difference in terms of variability between the two groups (i.e., the rats who were rewarded for varying their responses did so much more than did rats that were not reinforced for doing so). However, what is really interesting about the results of this study by Neuringer et al. (2001) is that the shift to extinction conditions resulted in a significantly reduced rate or responding for both subjects (evidence for the formation of an inhibitory assocation following the introduction of extnction conditions!) as well as an increase in the variability of responses made by both groups of subjects (Neuringer et al., 2001). Furthermore, as Domjan (2009) notes in his commentary on this study, the fact that these were the only effects is evidence that extinction "... did not alter the basic structure of the behavior," (Domjan, 2009, p. 302)- a fact which is in itself further evidence that extinction doesn't involve the unlearning of a previously acquired behavior, since, had that been the case, the effects of the introduction of extinction conditions would have been much more drastic!

Furthermore, Conklin and Tiffany (2002) identify several factors which pose challenges for research attempting to translate results from extinction studies involving animals into results applicable to human participants attempting to overcome their drug dependence. These include "... the renewal effect, spontaneous recovery, reinstatement, and failure to extinguish the most salient conditioned cues," (Conklin and Tiffany, 2002, p. 159).
Conklin and Tiffany (2002) begin their discussion of such challenges by describing the renewal effect by citing the fact that the context in which extinction was conducted can play a major role in determining the extent to which animals will maintain the same failure to respond they learned during extinction training when faced with a new context. To this ends, Conklin and Tiffany (2002) operationally define "contexts" as being the environmental locations in which drug-taking takes place,and the presence of cues which normally immediately precede a drug self-administration. As they go on to explain, contexts can play a crucial role in the expression of the learned behaviors involved in drug addiction; for instance, if a the behavior of repeated drug self-administration is acquired in one context and extinguished in another, a switch to a different context following extinction can often lead to a re-emergence of the previously-acquired drug-taking behavior! (Bouton and Ricker, 1994). Unfortunately, renewal often occurs when former drug addicts return to their original home location following the completion of their drug treatment, resulting in relapse (Conklin and Tiffany, 2002).
        Conklin and Tiffany (2002) go on to explain the mechanisms that might be responsible for the appearance of this effect. Apparently, "... following extinction, the CS has acquired two meanings: one associated with the original conditioning and one associated with extinction. The context in which the cue is presented determines which of these two meanings will be expressed," (Conklin and Tiffany, 2002, p. 160). This explanation underscores the fact that extinction involves the learning of another, inhibitory assocation, on top of the association that has originally been learned during acquisition! It also presents a significant challenge for researchers attempting to extinguish renewal effects among previous drug addicts.
       One means of decreasing the detrimental effects of renewal upon abstinence from drugs has been to attempt to extinguish a problem behavior in the same context as the one in which it was originally learned. While this does indeed result in decreased renewal effects, unfortunately, the decreased rate of renewal effects which results is often specific to the context in which the acquisition and subsequent extinction took place, and does not generalize to other contexts. A study by Bouton and Ricker (1994) involving rats confirmed this effect. In that study, Bouton and Ricker (1994) found that rats who had acquired a particular pattern of conditioned responding in one context and then had that pattern of responding extinguished in that same context, they did indeed fail to demonstrate renewal effects in that context. However, when, following this, these same rats were tested in different context, they readily showed renewal effects. Furthermore, Bouton and Ricker(1994) found this to be true in both experiments involving conditioned suppression and experiments involving appetitive conditioning. Bouton and Ricker(1994) also found these results to hold true when both contexts were equally familiar and equally associated with reward (or equally associated with both reward and lack thereof). As Bouton and Ricker (1994) went on to explain, and both Conklin and Tiffany (2002) and Domjan (2009) also commented, these results indicate that the effects of acquisition generalize more easily than do the effects of extinction, because the effects of extinction are more context-dependent. This presents bad news for therapies aimed at the extinction of conditioned behaviors.
       Luckily, however, there is a more effective means of preventing the renewal effect from occurring; a study by Gunther (1998), for instance, found that the effects of comparator stimuli could be dependent on context in the same way as could regular Pavlovian associations. As Conklin and Tiffany (2002) argue, this is one among several lines of evidence that when extinction is trained in several contexts, its effects generalize more readily than when it is trained in a single context (in other words, when extinction training is conducted in several contexts, the renewal effect is less likely to occur than when extinction training is conducted in only one context).
       As Conklin and Tiffany (2002) mention, this presents several interesting new questions, including one in regards to how many different contexts extinction should optimally be trained in, and whether the additive effects of conditioning extinction in different contexts ultimately plateau, and even whether the optimum number of contexts in which extinction should ideally be trained might vary between drugs.
       Interestingly, another study also conducted by Bouton and Brooks (2003) found that pairing a cue with the context in which extinction is conducted and then having that cue serve as a reminder of the extinction context can significantly reduce renewal effects. This finding led Bouton and Brooks (2003) to conclude that renewal effects often appear due to a failure upon the part of the subject to remember the context of extinction when they returned the context of acquisition  (i.e.,when they returned to the outside world after completing treatment at an inpatient facility for instance). This is consistent with Conklin and Tiffany's (2002) explanation (described above) of how, over the course of both acquisition and later extinction conditioning, the CSs associated with drug use (i.e., the needles used in heroin injections) acquire the capacity to elicit memories of both acquisition and extinction - and how, all too often, they stimulate memories of acquisition (as opposed to extinction) when former addicts return home following inpatient treatment.
Conklin and Tiffany (2002) go on to suggest several means of enhancing the effectiveness of such a cue, including those of having the reminder cue be novel and paired only with extinction (to avoid having it become associated with multiple processes, in the same manner that contextual cues, for instance, come to be associated with both acquisition and extinction), and, to the same ends, having it only be used in situations in which the former addict really needs a reminder of the extinction context (again, in oder to avoid a situation in which the reminder cue comes to be associated with too many different stimuli, resulting in its becoming less effective). Domjan (2009) presents several possible examples of such cues, including a calling card with the phone number of the patient's therapist, whom the patient is instructed to call if they feel that they are in danger of relapsing (in this situation, the calling card, unbenknownst to the patient, also serves the secondary function of calling forth memories of the extinction context!)

Conklin and Tiffany (2002) also discuss another phenomenon which can wreak havoc on results obtained during extinction training- namely, the phenomenon on spontaneous recovery, which, as they mention, can occur if a conditioned stimulus is presented after some time has been allowed to pass since the completion of extinction training.As Conklin and Tiffany (2002) explain, the crucial factor in spontaneous recovery is the passage of time between the end of extinction training and the time at which a subject is re-exposed to a conditioned stimulus.
Thus, Conklin and Tiffany (2002) argue that the effects of extinction can also be optimized via the manipulation of both the time intervals between sessions during which a participant is exposed to a cue to which extinction is being conditioned and the time intervals between the exposures to a particular cue within a particular session. To this ends, Conklin and Tiffany (2002) criticized many animal studies for conducting extinction via the presentation of multiple different types of cues within a single session, with each cue being presented only briefly, arguing that such brief presentations of each cue prevent any one cue from being extinguished completely. Instead, they argue that extinction should proceed with one cue at a time, with the target cue being presented multiple times within a single experimental session.
      Additionally, the results of a research conducted by Rescorla (2004) which involved  several experiments, including a classical conditioning procedure to train appetitive conditioning in rats, a separate experiment involving conditioning of instrumental responding in rats, and a sign-tracking experiment with pigeons found that the magnitude of spontaneous recovery is greater when shorter intervals of time are allowed between the intial acquisition training and later testing. These results stand in contrast to results obtained by a study performed by Myers, Ressler, and Davis (2006), on fear extinction. In that study, Myers et al. (2006) examined the effects of the spacing between acquisition trials and extinction trials on the extinction of fear. Myers et al. (2006) took as the basis for their study the understanding the acquisition of conditioned fear causes changes in the brain at the level of synaptic connections which can be "unlearned" (where "unlearned" is operationalized as the changes no longer remaining visible) a short time after the fear learning takes place, but not following the passage of a longer amount of time. Given these premises, they hypothesized that perhaps newly-acquired fears can be successfully extinguished (i.e., with the effects of spontaenous recovery, renewal, and reinstatement not taking hold) if extinction conditioning is performed shortly after the acquisition of a fear, but not if a longer period of time is allowed to pass. Their results (obtained from a behavioral experiment conducted with rats) were consistent with this hypothesis, showing that, in cases where extinction conditioning was conducted 24-72 hours after the original acquisition of fear, the subjects showed the expected patterns of renewal, spontaenous recovery, and reinstatement effects (all of which are challenges to successful extinction). By contrast, Myers et al. (2006) also found that, among the rats with whom extinction was conducted only between 10 and 60 minutes following initial acquisition, these effects were far less likely to occur. On the basis of these findings, Myers et al. (2006) concluded that, in order to increase the effectiveness of extinction procedures to optimum levels, it might be useful to conduct extinction procedures as soon as possible following acquisition. While the fact that these studies came out with such contradictory results might seem strange at first, Domjan (2009) effectively resolves this contradiction by explaining that the designs of the two studies were very different, with the study by Myers et al. (2006) being a within-subjects investigating fear conditioning, and the research by Rescorla (2004), by contrast, utilizing a between-subjects design to explore various types of appetitive conditioning.

In terms of translating results from animal studies such as those by Myers et al. (2006) and Rescorla (2004) into successful extinction treatment for people recovering from drug addictions, Conklin and Tiffany (2002) reccomend extinguishing one cue stimulus at a time, and allowing for the passage of sufficient amount of time for spontaneous recovery to occur, with the idea being that, once it occurs, extinction treatment would then be conducted once again, with the process repeating up until such a time as when responding to that particular cue is extinguished entirely. Furthermore, they recommend that after responding to a particular cue has been entirely extinguished further extinction training to that cue should still be conducted, following the passage of a significant amount of time. Finally, Conklin and Tiffany (2002) argue that the specific patterns according to which extinction training should ideally be conducted likely vary across subjects, and should be determined according to each individual's pattern of responding. They ackowledge, however, that until the specific mechanisms underlying extinction training are better understood, it will be difficult to put these recommendations into practice.

Conklin and Tiffany (2002) continue their discussion  cue-exposure addiction treatments by turning their attention to the phenomenon of reinstatement, which is another threat to the long-terms maintenance of extinction. Specifically, they define reinstatement as "... a phenomenon whereby responding to an extinguished CS re-emerges as a consequence of post-extinction exposures to a US," (Conklin and Tiffany, 2002, p. 162). They go on to explain that "After a Condtioned Stimulus has been extinguished, responding can be reinstated by presenting the Unconditioned Stimulus alone in the conditioning context. When the extinguished Conditioned Stimulus is subsequently presented alone in that context, conditioned responding can then occur as it did prior to extinction," (Conklin and Tiffany, 2002, p. 162).
In a study performed by Carroll and Comer (1996) for instance, a drug was administered to a group of animals. After a while, administrations of the drug were replaced with those of a vehicle (i.e., a CS for the drug) by itself. Over time, this resulted in the extinction of responding. Following this, the researchers administered various interoceptive and exteroceptive stimuli to the animals, in order to see if they would result in the animals making the same sort of responses as they did to the reinforcing effects of the drug itself. The study especially focused on the use of interoceptive stimuli, such as priming injections of the drug that had been previously used in the same study, as well as other drugs. The results of the study indicated that priming injections of a drug, especially a drug of the same class as the drug to which conditioned responding had previously developed during acquisition, (and later been extinguished during extinction), did indeed result in the the reinstatement of the animals' previous response. Interestingly, Carroll and Comer (1996) also noted that the magnitude of the reinforcement effects was directly proportional to the magnitude of the interoceptive stimuli administered, and that other factors- namely, restricted food intake and stress- also increased the magnitude of the reinstatement effects, at least studies of animals that had been previously addicted to cocaine. The results of this study are interesting both in that they offer another line of evidence [in addition to that offered by the study conducted by O'Brien et al. (1990) that exposing addicts to even a small amount of the drug to which they were previously addicted to can trigger relapse- and in that they suggest that other environmental factors, in addition to exposure to drug-related cues and contextual stimuli, can trigger relapse. Furthermore, the fact that, as Carroll and Comer (1996) mention, relapse can also be triggered by stress and inadequate nutrition seems to imply that perhaps individuals newly exiting rehabilitation facilities should be provided with a gradual reintroduction to their previous living environment and responsibilities, so as to preclude the possibility of the stress of the transition causing them to relapse.
       In regards to the possibility of reinstatement effects triggering relapse following extinction of drug-taking behaviors among newly-recovering addicts, Conklin and Tiffany (2002) mention that "Although it may not be common for addicts to experience non-contingent re-exposure to illegal drugs, it is common for addicts in recovery to be exposed to drugs such as painkillers or cough medicines that may contain narcotics or alcohol," (Conklin and Tiffany, 2002, p. 162). These incidental exposures, combined with possible lapses in the addicts' own willpower and the breaches in abstinence that may result in the reinstatement of conditioned responding to drug cues. As a result of this and other possibilities for relapse following the extinction of drug self-admnistration behaviors, Conklin and Tiffany (2002) suggest, based on research conducted with animals, that one of the most effective means of enhancing the robustness of extinction treatments, in terms of preventing relapse, may be to combine them with other forms of psychotherapy.


Bouton. , & Brooks, (1993). Time and context effects on performance in a pavlovian discrimination reversal. Journal of Experimental Psychology: Animal Behavior Processes., 19(2), 165-179. doi: 10.1037/0097-7403.19.2.165

Bouton, M. E., & Ricker, S. T. (1994). Renewal of extinguished responding in a second context. Animal Learning & Behavior., 22(3), 317-324. doi: 10.3758/BF03209840

Carroll, M. E., & Comer, S. D. (1996). Animal models of relapse. Experimental and Clinical Psychopharmacology, 4(1), 11-18. doi: 10.1037/1064-1297.4.1.11

Conklin, C. A., & Tiffany, S. T. (2002). Applying extinction research and theory to cue-exposure addiction treatments. Addiction, 97(2), 155-167. doi: 10.1046/j.1360-0443.2002.00014.x

Domjan, M. (2009). Learning and behavior. (6 ed., pp. 107, 115). Belmont, CA: Wadsworth, Cengage Learning.

 Ehrman, R. N., Robbins, S. J., Childress, A. R., & O'brien, C. P. (1992). Conditioned responses to cocaine-related stimuli in cocaine abuse patients. Psychopharmacology, 107(4), 523-529. doi: 10.1007/BF02245266

Gunther, L. M., Denniston, J. C., & Miller, R. R. (1998). Renewal of comparator stimuli. Learning and Motivation, 29(2), 200-219. doi: 10.1006/lmot.1998.1003

Hammersley, R. (1992). Cue exposure and learning theory. Addictive Behaviors, 17(3), 297-300. doi: 10.1016/0306-4603(92)90035-T

Myers, K. M., Ressler, K. J., & Davis, M. (2006). Different mechanisms of fear extinction dependent on length of time since fear acquisition. Learning & Memory, 13(2), 216-223. doi: 10.1101/lm.119806

Neuringer, A., Kornell, N., & Olufs, M. (2001). Stability and variability in extinction. Journal of Experimental Psychology: Animal Behavior Processes, 27(1), 79-94. doi: 10.1037/0097-7403.27.1.79

O'Brien, C. P., Childress, A. R., McLellan, T., & Ehrman, R. (1990). Integrating systematic cue exposure with standard treatment in recovering drug dependent patients. Addictive Behaviors, 15(4), 355-365. doi: 10.1016/0306-4603(90)90045-Y

Rescorla, R. A. (2004). Spontaneous recovery varies inversely with the training-extinction interval. Learning & Behavior., 32(4), 401-408. doi: 10.3758/BF03196037

Siegel, S., Baptista, M. A. S., Kim, J. A., & Weise-Kelley, L. (2000). Pavlovial psychopharmacology: The associative basis of tolerance. Clinical Psychopharmacology. Special Issue: The Decade of Behavior: Psychopharamcology and Substance Abuse Research, 8(3), 276-293. doi: 10.1037/1064-1297.8.3.276

Saturday, May 26, 2012

       In their article "Addiction," Robinson and Berridge (2003) explore the phenomenon of drug addiction from the perspective of seeking to answer the question of why it is that while many people experiment with drugs, only some become "addicts" (where "becoming an addict" is operationalized as developing a pattern of compulsive drug-seeking and self-administration which often comes at the expense of other activities). In particular, the seek to discover first, what it is about certain individuals which makes them particularly susceptible to the development of such habits, and second, what it is that makes it so difficult for these individuals to overcome their drug-taking habits once they are established.

       In attempting to answer these question, Robinson and Berridge (2003) begin by suggesting that the potent and addictive effects of drugs rely upon their ability to "hijack" the brain systems normally involved in the processing of learning about rewards and in enabling organisms to attribute motivational value to stimuli important to survival (Kelley and Berridge, 2002). They also argue that successive drug administrations can result in changes in these neural systems which can lead the drug-taking behavior becoming increasingly compulsory. Among the neural systems involved in these processes, they include those of the "... dopamine projections from the ventral tegmental area and the substantia nigra to the nucleus accumbens (NAcc) and striatum, as well as glutamate inputs from the prefrontal cortex, amygdala, and hippocampus and other key parts of the system to which [they] refer to as NAcc-related circuitry," (Robbinson and Berridge, 2003, p. 26).  According to Robinson and Berridge (2003), the connection between these systems and the process of addiction is rooted in first, the fact that drugs are able to engage these neural systems to a greater extent and more powerfully than can natural rewards, and second, the fact that over successive administrations, drugs can actually change the way in which these systems operate, in such a way as to make drug-taking behavior increasingly compulsory, (Robinson and Berridge, 1993, 2000).

Robinson and Berridge (2003) go on to present several possible accounts for what these drug-induced changes in psychological functioning might be, and of how they might come about as a result of drug-induced changes in brain circuitry.

         One of these involves the Opponent Processes Theory of Motivation (Solomon & Corbit 1973), in which a compensatory b-process (a process designed to return the system back to a state of homeostasis following a drug-induced disruption), which originally comes on only as a gradual decay in the intense reward state induced by drug self-administration, begins, over time, to come on sooner and be stronger when it does so. Over time, the b-process begins to come on earlier and to be stronger when it does so. This phenomenon of the b-process becoming stronger eventually allows it to compete successfully with the a-process (the set of physiological processes triggered by the ingestion of the drugs themselves- i.e., the rapid heart rate and feelings of euphoria which follow the ingestion of cocaine, for example). The result of this "competition" between the stimulus-bound, drug-induced a-process and the compensatory b-process is that the effects of the a-process come to be greatly attenuated (i.e., the addict obtains progressively less of the desired effects from each successive drug self-administration). Furthermore, the now-stronger b-process causes progressively more severe and debilitating withdrawal symptoms (the withdrawal symptoms come on as a result of the fact that, over time, the b-process becomes able to not only to counteract the effects of the a-process, but also to outlast its effects). Crucially, only the effects of the b-process change over time. The effects of the stimulus-dependent a-process neither increase not decrease over time; rather, they become increasinly attenuated solely as a result of the increasing ability of the b-process to counteract  them. The withdrawal symptoms tend to be particularly severe when the addict is exposed to the types of environments and other conditioned stimuli in the presence of which they would usually have taken the drugs, but is deprived of the opportunity to self-adminster the drugs (Siegel et al. 2000). At the same time, the increasing attenuation of the a-process results in the addict having to take ever-increasing amounts of the drug in order to obtain the same degree of desired effects as they had been able to before (the phenomenon of the buildup of "drug tolerance"). The overall result of these phenomena is the seemingly-paradoxical phenomenon of addicts continuously self-administering drugs, to the detriment of their health and personal well-being, not for the purpose of attaining any sort of pleasurable effects, but rather solely with the goal of avoiding the debilitating withdrawal symptoms which are certain to ensue should they cease to do so! In this way, as Robinson and Berridge argue, drug taking can become truly compulsory (Robinson and Berridge, 2003). Interestingly, according to Robinson and Berridge (2003), "Once the b-process is strengthened, even a small drug dose can instate it and thereby trigger withdrawal again. Conversely, prolonged absitnence from the drug would decay the b-process, and the ability to reactivate it would return back to normal. Once the b-process returns back to normal, the person would no longer be addicted," (Robinson and Berridge, 2003, p. 28). This presents a strong argument for treatment programs which emphasize total abstinence from alcohol. It also goes some distance in explaining the phenomenon of people who recently graduated from detoxification programs accidentally overdosing on their "one last time," intake of their drug of choice.

       In their explanation of how the Opponent Processes Theory of Motivation might work at the neural level,  Robinson and Berridge (2003) suggest that "... the positive a-process is caused by activation of mesolymbic dopamine projections to the nucleus accumbens and amygdala which in turn govern the reinforcing properties of drugs," (Robinson and Berridge, 2003, p.29) (Koob and Le Moal, 1997). Furthermore, they propose that the buildup of tolerance that comes as a result of repeated drug self-administrations comes as a consequence of a "... downregulation of in the mesolimbic dopamine system," (Robinson and Berridge, 2003, p. 29) which repeated drug self-administration causes. Furthermore, they note that sudden cessation of drug intake (i.e., as might come about when people quit "cold turkey"),  "... causes dopamine (and serotonin) levels to drop below normal levels... resulting in a dysphoric B-state of withdrawal," (Robinson and Berridge, 2003, p. 29). As if this were not enough, sudden cessation of drug self-administration apparently also causes the activation of "...additional compensatory b-processes via the hypothalamic-pituitary axis stress system, causing the release of corticotropin releasing factor (CRF) in the amygdala, as well as other stress responses," (Robinson and Berridge, 2003, p. 29) (Koob and Le Moal, 1997). As these researchers argue, such effects cause addicts to what has quickly become (as a result of these neurochemical changes) a negative emotional and motivational state. In order to escape the aversive nature of this state, addicts may attempt to "self-medicate" by resuming to take at least some amount of the drug to which they were previously addicted, rekindling their previous dependence upon it.

     Robinson and Berridge (2003) go on to offer several interesting criticisms of the Opponent Processes Theory of Motivation as applied toward explaining drug addiction. For one thing (they argue), withdrawal (induced by activation of the b-process, as noted above), may actually be a far less powerful motivator of drug-seeking behavior than either the pleasurable effects which come as a direct result of taking the drugs, or the effects of stress, (Stewart and Wise, 1992)  For example, Stewart and Wise (1992) performed a study involving rats and lever-pressing behavior as part of a study attempting to elucidate which factors might be responsible for stimulating relapse among rats which had previously been addicted to drugs (cocaine or heroin specifically) but who had since been deprived of the drugs for a sufficient amount of time to break their addiction. In that study, Stewart, Shaham, and other researchers measured the amount of lever-pressing which the rats were willing to perform under extinction conditions in order to obtain the reward of an injection of the drug to which they had previously been addicted (Stewart and Wise, 1992). In that study, the researchers performed the manipulation of activating either the "a process," or the "b process." "Activating the 'a' process was done by giving the rats a small injection of the drug to which they had previously been addicted (Stewart and Wise 1992).To activate the b-process, the researchers administered a drug called naltrexone to the rats (Stewart and Wise, 1992). Naltrexone is an opioid antagonist, which means that it blocks the brain's naturally-occurring opioid receptors, with the result that when it is administered to individuals who are dependent on the opioid drug heroin, these individuals begin experiencing withdrawal symtoms (Stewart and Wise, 1992). For a heroin addict, the state which they would be experiencing following an adminstration of naltrexone is similar to the type of drug-withdrawal state which they would be experiencing if they suddenly ceased self-administering heroin (Stewart and Wise, 1992). As Robinson and Berridge (2003) explain, in the context of the Opponent Processes Theory of Motivation as applied to drug addiction, the artificial "withdrawal" state which an administration of naltrexone induces among former drug addicts is likely to be the most powerful force for getting them to relapse and resume self-administering heroin (this is the case because, according to the Opponent Processes Theory of Motivation's explanation of drug addiction and relapse, the withdrawal symptoms which individuals experience following their ceasing to self-administer a drug are the most powerful force toward motivating them to relapse in their drug-taking habits. Interestingly, however, the results of the Stewart and Wise (2003) experiment would suggest that, actually, both the administration of a small amount of the drug of choice, or exposure to a stressor are actually much more powerful means by which to induce a relapse in drug-seeking behavior, than the naltroxene-induced experience of withdrawal symptoms! (Stewart and Wise 1992). In other words, activation of the A process (as operationalized by exposing the rats to a small amount of the drug to which they had previously been addicted), is actually much more powerful than activation of B-process withdrawal symptoms. This would imply that there must be another force at work, aside from the experience of the withdrawal symptoms alone, which would be motivating individuals to continue self-administering the drugs.

As Robinson and Berridge (2003) also note, another problem which the Opponent Processes Thoery of Motivation has difficulty explaining is that of the unfortunate case in which individuals who were previously addicted to drugs, but who have since been abstinent for a significant period of time suddenly begin self-administering drugs, seemingly of their own volition. This especially seems to be the case when one notes that, by this time, the b-process (which will have been causing negative withdrawal symptoms) should have decayed away a long time ago- and, thus, according to the Opponent Processes Theory of Motivation, these individuals should no longer be addicted to drugs!
A possible explanation for this, as Domjan notes, is that of the drug cues surrounding the  self-administration of a drug (i.e., being in a particular location, being surrounded by particular friends, etc.) becoming, over many consecutive drug self-administrations, conditioned to act as conditioned stimuli predicting the impending self-administration of drugs. This also allows them to stimulate the initiation of compensatory b-processes, allowing an individual's body to compensate ahead of time for the dose of the drug which the conditioned cues predict. In accordance with this, studies with both humans and animals have shown that indeed,  such conditioned cues can, indeed, elicit withdrawal symptoms- at least in theory (Robinson and Berridge, 2003). The problem, however, is that they often do not; in fact, as Robinson and Berridge (2003) report, many former addicts report that they fail to do so! The fact that conditioned cues related to the self-administration of drugs do not appear powerful enough, on their own, to explain why relapse occurs has led researchers to search for different explanations.

       One such alternative, which Robinson and Berridge (2003) also present, is the possibility is that addiction to drugs comes as the result of learning processes. According to Robinson and Berridge (2003),this view came about as a result of a series of findings illustrating the role of neural circuitry, particularly that associated with the nucleus accumbens, in learning about rewards. Most hypotheses attempting to explain drug addiction through the role of learning processes argue that what occurrs specifically is "... drugs produce abnormally strong or aberrant associations involved in reward learning, more powerful than natural reward associations," (Robinson and Berridge, 2003). According to Robinson and Berridge (2003), the specific type of learning occuring in these types of situations could be of pretty much any kind.

The first possibility which Robinson and Berridge (2003) consider is that of drugs being involved in explicit learning, in which the association formed just happens to be abnormally strong. As Robinson and Berridge (2003) state, in such a case, individuals addicted to drugs will simply have formed an unusually strong action-outcome association between the action of self-administering drugs and the outcome of receiving them, as well as another, also unusually strong, association between perceiving certain environmental cues and the imepending delivery of drugs (Balleine and Dickinson, 1998).
       Evidence for the fact that humans establish Response-Outcome associations in this manner comes, for instance, from a study by Balleine and Dickinson (1998) establishing the existance of what they call a "contingency degradation effect"- the idea that reducing the contingency between an action and a desired outcome by adding extra instances of the delivery of the outcome (reinforcer) should (and does) reduce the number of instances of the action.
      Similarly, a study by Colwill and Resorla (1986) established the existance of what they called the "Reinforcer Devaluation Effect"- the idea of which is that, if a subject performs an action in order to get a desired effect, then reducing the desirability of the effect should lead to a decrease in the numnber of instances of the action preceding it (which is exactly what happens).
       The fact that Robinson and Berridge (2003) include animals, as well as people, in this category of individuals who can form such abnormally-strong associations
differentiates their work from, for example, that of Thorndike, who, as Domjan states, all throughout his studies involving animals learning to escape out of puzzle boxes and the like, remained convinced that the animals' escapes were merely the result of trial-and-error learning, rather than their becoming aware, in any kind of intelligent way, of their specific actions to escape the box and the ensueing positive outcome. Rather, he believed the animals initially made their escapes through trial-and-error, and then simply repeated the same type of escape over and over again.
        According to Robinson and Berridge (2003), "Abnormally strong explicit learning might distort declarative memories or expectations in two ways. (a) Conscious memories of the hedonic drug experience might be especially vivid and/or abnormally intrusive. (b) Drugs could exaggerate or distort declarative memories such that memory-based cognitive expectations about drugs become excessively optimistic," (Robinson and Berridge, 2003, p. 32). As Robinson and Berridge (2003) go on to explain, however, addicts' excessive drug-taking behaviors do not seem to be tied to distorted memories and excessively positive expectations about the consequences of doing so. Evidence for this comes through such sources as what the addicts themselves say about their lives and the amount of pleasure they expect from drug taking- which they themselves agree does not justify the excessive costs of  continuing to self-administer controlled substances (Robinson and Berridge, 2003).
These conclusions are also consistent with the results of the previously-mentioned study by Colwill & Rescorla (1986) establishing the existance of the reinforcer devaluation effect- the main conclusion of which was that devaluing the value of a desired reinforcer should lead a decrease in the occurences of actions being performed with the goal of obtaining that reinforcer. Since, as Robinson and Berridge (2003) claimed, most addicts do not have unrealistically positive expectations of regarding the types of outcomes which drug-seeking and drug-taking behaviors, it appears that the "outcome" in this situation (i.e., whatever sensations they obtain from taking the drugs that they do) has effectively been "devalued" for them. Given this, according to the results of the study by Colwell and Rescorla (1986), the drug-seeking behaviors should lessen or cease. Thus, it makes sense that Robinson and Berridge (2003) claim that drug addiction likely cannot be explained as being rooted in the formation of conscious, although overly-intense and overly-positive, response-outcome associations between the response of drug self-administrations and the outcome of drug-induced "highs."

In response to this, Robinson and Berridge (2003) came up with an alternative explanation for drug addiction- that of it being the result of the establishment of deeply-ingrained and implicit (rather than declarative) S-R (stimulus-response) or S-S (stimulus-stimulus) habit associations within an addict's mind. As Robinson and Berridge (2003) note, the most promising of these possibilities is that of addicts' drug habits progressing from what are initially readily-verbalized, declarative-memory associations between the action of taking a drug (a response) and the outcome of a "high,"which, based on previous congitive expectations has been labeled a "desirable outcome," to to more automatic, stimulus-response guided behavior. As habits, the latter associations occur without the involvement of conscious memory processes. Essentially, such hypotheses suggest, that, as was suggested in my Fundamentals of Learning class lecture, over-learned action patterns often become automatic habits. What Robinson and Berridge (2003) add to this is the idea that, over time, such habits can actually become compulsive.
This idea also makes sense in light of a study conducted by Adams (1982), which sought to investigate the question of why behavior is mediated by goal-directed, stimulus-response associations at some times, and by habitual, stimulus-response assocations at other times. In that study, rats were initially trained to press a lever (make a response) in order to obtain the reward of one sucrose pellet (an outcome/reinforcer) (Adams 1982). Initially, the schedule of reinforcement in this study was one of "continuous reinforcement," or a "fixed response or Fr1 schedule"- that is, the rats were rewarded with a sucrose pellet for every lever press they made (Adams 1982). The rats were then divided into two groups: one group received 100 trials of training, while another group received 500 trials of training (Adams 1982). Then, the rats in each of these two groups were further divided up into two groups (i.e., the rats in the "500 trial group" were divided into two groups, as were the rats in the "100 trial group")(Adams 1982). For one of the two groups of rats in each of the trail conditions (i.e., the 500-trial group, and the 100-trial group, respectively), the sucrose-pellet outcome was then devalued though pairing the administration of the sucrose-pellet outcome with the administration of an aversive lithium chloride stimulus (Adams 1982). For the other half of the rats in each number-of-tirials condition, the stimulus was not devalued (i.e., the sucrose pellets were still delivered normally(Adams 1982).Then, both groups of rats were tested under extinction conditions in order to see how many lever pressess they would make (Adams 1982). Fascinatingly, for the rats in the 100-trial group, the lever-pressing behavior remained goal-directed; that is, the rats would only press the lever if doing so allowed them to attain a desired outcome (that of being rewarded with regular sucrose pellets, which they saw as desirable, and not lithium-chloride sucrose pellets, which they saw as undesirable)(Adams 1982). Among the rats in the "500 trail group," however, all rats continued to press the lever, whether or not the sucrose-pellet outcome for doing so had been devalued! (Adams 1982). The results of this experiment by Adams (1982) lend further support to the idea that overtraining a behavior can cause it to transition from being a goal-directed behavior to becoming one that is habitual and mindlessly insensitive to the outcome it might result in.

However, in response to this, Robinson and Berridge (2003) offer the interesting criticism that while an overtrained behavior is indeed likely to become habitual, in their words, "... habits are not necessarily likely to become compulsive in any motivational sense, no matter how automatic they are,"(Robinson and Berridge, 2003, p. 33). To this ends, Robinson and Berridge (2003) present the examples of habits such as brushing one's teeth and the like and argue that, while these are habits and thus require minimal cognitive attention, they are not motivationally compulsive; people, with very few exceptions,  feel absolutely compelled to do them, even in the face of great sacrifice. Indeed, Robinson and Berridge (2003) argue that the sheer flexibility and the great range of behaviors in which addicts are willing to engage in order to obtain the drugs they crave, as well as the severely agitated and distressed manner in which many addicts react when they are suddenly deprived of a routine dose of the drugs to which they have become accustomed mandates a search for further explanations of drug addiction, beyond simply the formation of strong habits.

Similarly, while Robinson and Berridge (2003) briefly consider the possibility that perhaps drug addiction is mediated by a distorted form of learning in which an addict might form an unreasonably strong association between, for instance, a particular location and the environmental stimuli present there, and the administration of a drug, they ultimately dismiss the possibility that such actions might be able to elicit compulsive drug taking. Rather, they suggest that "... most S-S associations may actually remian normal in addicts. What is aberrant in aberrant in addiction is the response of brain motivational systems to Pavlovian-conditioned drug cues," (Robinson and Berridge, 2003, p. 35).

Thus, in sum, Robinson and Berridge (2003) argue that ".. learning and incentive motivational processes are joined in the transition to addiction," (Robinson and Berridge, 2003, p. 36). Their view that learning and motivational processes are joined in the development of addictions is evident in their own theory of drug addiction, which they refer to as that of "incentive sensitization," (Robinson and Berridge, 2003, p.36).

 According to the "incentive sensitization," (Robinson and Berridge, 2003, p. 36) view of addiction pioneered by Robinson and Berridge (2003), drug addiction involves "drug-cues triggering excessive incentive motivation for drugs, leading to compulsive drug seeking, drug taking, and relapse," (Robinson and Berridge, 2003, p. 35), (Robinson and Berridge, 1993, 2000).The main idea behind this theory is that the self-administration of drugs creates lasting changes in the brain's nucleus accumbens-related structures. The structures which drugs tend to modify are those which attribute salience (or lack thereof) to various environmental stimuli. As a consequence of these drug-induced modifications, areas in the nucleus accumbens become lastingly hypersenstive [or as Robinson and Berridge call it, "sensitized," (Robinson and Berridge, 2003, p. 36) to the effects of particular drugs and to the stimuli predicting their administration. These changes lead areas in the3 nucleus accumbens area of the brain to attribute excessively-high ratings of stimulus salience to drugs and cues related to their self-adminstration- a process which, on a psychological level, leads to extreme and abnormal desire to take drugs. As Robinson and Berridge (2003) go on to explain, such abnormal "wanting" "...can sometimes become manifest implicitly in drug-seeking behavior," (Robinson and Berridge, 2003, p. 36). The fact that this often happens at an implicit level means that individuals often remain unaware, on a conscious level, of the processes taking place [although Robinson and Berridge (2003) go on to explain that sometimes other cognitive processes can sometimes cause these representations which these processes involve to become consciously manifest, resulting in the affected individuals developing a very consciously-noticable desire to take drugs.
       Significantly, Robinson and Berridge (2003) draw a distinction between "the sensitized neural systems responsible for incentive salience," (Robinson and Berridge, 2003, p. 36)- in other words, those responsible for how much a particular drug might be "wanted"- and "the neural systems that mediate the hedonic effects of drugs, how much they are 'liked,'" (Robinson and Berridge, 2003, p. 36). They further suggest that these two processes are governed by different areas of the brain, and thus largely operate independently of each other.
       In defining "sensitization," Robinson and Berridge (2003), speak of it being largely the opposite of tolerance; that is, in sensitization, with each subsequent self-administration of a drug actually increases its effects. According to Robinson and Berridge (2003), there are two major classes of sensitization: psychomotor sensitization, and motivational sensitization. Both, however,are sensitive to changes in the neural structure of the nucleus accumbens- a process which they argue takes place when individuals repeatedly self-administer drugs. Interestingly, psychomotor sensitization, at least, is context-specific; for instance, rats which had been exposed to stimulants which resulted in their exhibiting symptoms of psychomotor sensitization only showed these symptoms when they were subsequently tested in an environment that contained many of the same cues as had been present when the drugs were orginally being administered to them (Robinson and Berridge, 2003).
       Interestingly, as Robinson and Berridge (2003) note, individuals can differ in their degree of susceptibility to sensitization- something which they argue can be helpful in unraveling the puzzle of why some individuals, but not others, become drug addicts. Furthermore, they note that "... once sensitized, most individuals show cross-sensitization, which means that sensitization to one drug can cause sensitized effects for other drugs as well," (Robinson and Berridge, 2003, p. 38).  Finally, Robinson and Berridge (2003) note that cross-sensitization can even occur between drugs and nondrug stress! Taken together, these factors begin to suggest how such factors as gateway drugs and exposure to stress may come into play in addiction.

In terms of the specific changes in the brain involved in sensitization, Robinson and Berridge (2003) argue that "sensitization-related changes have been described in many neurotransmitter systems that are integral to the function of NA-cc related circuits including serotonin, norepinephrine, acetylcholine, opioid and GABA systems," (Robbinson and Berridge, 2003, p. 38). Thus, the changes in the nucleus accumbens by sensitization appear to be quite widespread. Furthermore, drug-induced changes in the nucleus accumbens occasionally occur even at the length of neurons, with neurons in the nucleus accumbens and in the prefrontal cortex often showing "...changes in the lengths of dendrites and the extent to which dendrites are branched," (Robbinson and berridge, 2003, p. 38). Furthermore, "...changes can also occur in the density and types of dendritic spines, which are primary site of excitatory gluatamate synapses," (Robinson and Berridge, 2003, p. 39). As Robinson and Berridge (2003) argue, such changes may significantly alter the way information about rewards is processed in this area of the brain.

Overall, these effects can  go some distance in explaining not only why it is that some people become addicted to drugs, but also why many of these individuals continue self-administering them, even in the face of deriving little pleasure and, often, much suffering as a result of continuing to do so.


Adams, C. D. (1982). Variation in the sensitivity of instrumental responding to reinforcer devaluation. The Quarterly Journal of Experimental Psychology B: Comparative and Physiological Psychology, 34B(2), 77-98. Retrieved from

Colwill, R. M., & Recorla, R. A. (1986). Postconditioning devaluation of a reinforcer affects instrumental responding. Journal of Experimental Psychology: Animal Behavior Processes, 11(1), 120-132. Retrieved from

Domjan, M. (2009). Learning and behavior. (6 ed., pp. 107, 115). Belmont, CA: Wadsworth, Cengage Learning.

Kelley, A. E., & Berridge, K. C. (2002). The neuroscience of natural rewards: Relevance to addictive drugs. Journal of Neursocience, 22(9), 3306-3311. Retrieved from Kelley, A. E., & Berridge, K. C. (2002). The neuroscience of natural rewards: Relevance to addictive drugs. Journal of Neursocience, 22(9), 3306-3311.

Koob, G. F., & Le Moal, M. (1997). Drug abuse: Hedonic homeostatic dysregulation. Science, 278(5335), 52-58. doi: 10.1126/science.278.5335.52

Robinson, T. E., & Berridge, K. C. (2003). Addiction. Annual Review of Psychology, 54, 25-53. doi: 10.1146/annurev.psych.54.101601.145237

Robinson, T. E., & Berridge, K. C. (2000). The psychology and neurobiology of addiction: An incentive-sensitization view. Addiction, 95(Suppl2), S91-S117. doi: 10.1080/09652140050111681

Robinson, T. E., & Berridge, K. C. (1993). The neural basis of drug craving: an incentive-sensitization theory of addiction. Brain Research Review, 18(3), 247-291. doi: 10.1016/0165-0173(93)90013-P
Siegel, S., Baptista, M. A. S., Kim, J. A., & Weise-Kelley, L. (2000). Pavlovial psychopharmacology: The associative basis of tolerance. Clinical Psychopharmacology. Special Issue: The Decade of Behavior: Psychopharamcology and Substance Abuse Research, 8(3), 276-293. doi: 10.1037/1064-1297.8.3.276

Solomon, R. L., & Corbit, J. D. (1973). An opponent process theory of motivation: I. temporal dynamics of affect. Psychological Review, 81(2), 119-145. doi: 10.1037/h0036128

Stewart, J., & Wise, R. A. (1992). Reinstatement of heroin self-administration habits: Morphine

prompts and naltrexone discourages renewed responding after extinction. Pychopharmacology, 108

(1-2), 79-84. doi: 10.1007/BF02245289