One of the most difficult challenges to overcome in life is getting out from under the grip of an addiction, whether it be drug, alcohol or nicotine dependency, a food addiction or eating disorder, or compulsive activities such as gambling, shopping, pornography or Internet addiction. Taken to the extreme, addictions can become highly self-destructive, antisocial or criminal activities such as self-mutilation, kleptomania, or pyromania. At the other end of the scale are ordinary activities, such as exercise or work, which in normal degree are healthful but when excessive can become addictive. There are also minor compulsions which might best be considered bad habits rather than addictions, such as nail biting, hair pulling and the like. Broadly speaking, an addiction can be any habitual behavior which takes over one’s life, interferes with social relations and personal achievement, and threatens one’s autonomy. There are many ideas about what addiction is and how to treat it, but unfortunately success rates are low and relapse rates are high. However, there is a recent approach to snuffing out addiction based on the emerging sciences of neuroplasticity and behavior modification, which holds out the promise of lasting change. The approach is called cue exposure theory, and it goes against the conventional wisdom. I will discuss it after first reviewing the more conventional approaches. And I’m going to do something else unusual at the end of this particular blog post: I will apply this methodology to an “addiction” of my own and follow my progress in the Discussion Forum associated with this blog.
Models of addiction. There are a number of different views of what addiction is. The medical model views addiction as a disease, focusing on the biological aspects of physical or psychological dependency. This view typically confines the idea of addiction to cases of substance abuse and dependency, and attempts to pinpoint the basis for addiction in terms of changes in brain circuitry and the chemical action of reward neurotransmitters such as dopamine and serotonin. The medical model also highlights the biological reality of withdrawal symptoms when the addictive substance is removed. A second model, the psychiatric model, looks at addiction somewhat more broadly as a manifestation of unresolved psychosocial or emotional conflicts that lead to compulsions or poor impulse control; often this is broadened to include the family, social or cultural context. A third model, which we might call the autonomy model, rejects the medical and psychiatric models as too deterministic and incompatible with the existence of free will. This model takes addiction to be fundamentally a question of personal responsibility and choice. Finally, behavioral models do not necessarily take a position on the origins of addiction, but look instead at how addictive behaviors can be modified and eliminated. Of course, there are many variations and combinations of these models of addiction.
Varieties of treatment. Depending on which model is favored, different treatments variously emphasize medical detoxification and the use of pharmaceuticals; individual, family, group or residential rehabilitation counseling; recognition of personal responsibility; or various modalities of behavior modification. Under the medical model, pharmaceuticals are often prescribed for detoxification and the relief of cravings. While drugs may in fact help reduce cravings in the short term, they can create their own problems of side effects and substitute addictions. Antagonist drugs, which block receptors for “reward” transmitters such as dopamine, are often unpleasant and create incentives to quit or circumvent treatment, and they invite relapse once they are discontinued. Typical success rates for drug and alcohol detox rehab programs, which combine medical detox and psychological or psychiatric treatment, have been cited to be as low as 2-20 percent. One such program, Narconon, claims a success rate of 76%, but this figure has been challenged as being vastly overinflated and based upon methodologically flawed statistics. As with many similar programs, Narconon insists on the importance of getting treatment away from the normal work-home environment: :
One thing is for sure if you are trying to break a habit such as drug addiction, a change of environment should be at the top of the list as far as solutions. Due to these factors, attending a drug rehab close to home is seldom the correct treatment option for chronic drug abusers. It is extremely therapeutic to be distanced from the people they used drugs with, drug dealers, and the surroundings that can continue to stimulate their past addictive behaviors.
As we’ll see shortly, it is precisely this key assumption that is questioned by cue exposure therapies.
Behavioral therapies. In essence, behavioral approaches look at addictions primarily as conditioned behavioral patterns that are strongly reinforced, but from which the addict nevertheless still has some motivation to escape. Behavioral therapies tend to divide into two camps: those which employ classical and operant conditioning to directly modify behavior by changing the reinforcement patterns; and those which supplement the conditioning techniques, or replace them entirely, with a cognitive element, following the model of Cognitive Behavioral Therapy (CBT). The cognitive element typically involves actively thinking about ones behavior, and reflecting on whether or not it is based upon rational or empirically valid assumptions. For example, CBT may treat depression, anxiety, or phobias by challenging an individual to consider whether one’s worst fears are in fact likely to happen, what one is giving up by maintaining the present behavior, and what one stands to gain by stopping it. Often meditation, mindfulness, and notions of self-efficacy are involved in these cognitive approaches. Examples of the application of CBT to addiction are Alan Marlatt’s Relapse Prevention Therapy and also his Mindfulness therapy; and Aaron Beck’s Cognitive Therapy of Substance Abuse.
However, overcoming addiction may not be all that susceptible to “reasoning” and reflection. Addictive cravings are often incredibly powerful and tend to overwhelm rational thinking.
Cue exposure therapies. There are two very different approaches to treating addiction by behavior modification: stimulus avoidance and cue-exposure therapies. While they are both considered “behavioral” treatments, they are in fact polar opposites! The stimulus avoidance therapies involve training the individual to avoid exposure to the stimulus. In practical terms, this means abstinence. It is the approach taken, for example, by Alcoholics Anonymous. A core assumption of AA is: “Once an alcoholic, always an alcoholic”. Those who take this view claim that it is impossible, or highly risky, for an alcoholic ever to return to moderate drinking. AA has a good success rate, but it tends to require a strong “spiritual” commitment, and can be sabotaged by relapse if the recovering alcoholic or addicts takes even a single drink.
There is an emerging area of research, however, which takes issue with the stimulus avoidance school of thought, and supports the idea that addictions can be replaced by normal responses to behavioral cues, using cue exposure therapy, sometimes called response prevention therapy. And even more radically, the treatment works best if carried out in the most realistic context of the daily life patterns of the addict. This completely contradicts the central assumption of Narconon in the above quote!
For a full explanation of the psychological basis and technical terminology of reinforcement theory, I would recommend reading the Psychology page of this blog, which provides useful background on the work of Pavlov and current applications by behaviorists such as Daniels and Pryor in the use of cue exposure as a general method for extinguishing behaviors. In short, the essence of cue exposure therapy is to extinguish the addictive behavior by allowing the addict to be exposed to normal cues or stimuli that typically precede the addictive behavior, but preventing that behavior from getting underway. This clearly leads to significant discomfort and even withdrawal symptoms in serious cases. However if repeated frequently enough, and in the presence of a sufficient variety of cues and contexts, cue exposure therapy can be very successful in extinguishing addiction. Even more importantly, there is evidence that is is successful in preventing relapse over the longer term.
Furthermore, cue exposure therapy is a general approach to addiction treatment. It works not only in treating “chemical” addictions of substance abuse, but addictive behaviors more generally. There are studies showing its effectiveness with treatment of drug and alcohol addiction, tobacco addiction, and eating disorders. For example, using cue exposure and response prevention, combined with gradualism may be more effective than going “cold turkey” for learning to permanently stop smoking. Other studies show that cue exposure therapy is more effective than a “self control” based cognitive behavioral approach in treating bulimia.
What makes cue exposure succeed? Despite encouraging data of the effectiveness of cue exposure therapies in both addiction cessation and relapse prevention, it is not always successful. A recent review article in the journal Addiction, by Conklin and Tiffany of Purdue, provides an excellent meta-analysis of 18 cue exposure therapy for treating a range of addictions–including treatments for addiction to alcohol (N=5), nicotine (N=5), cocaine (N=1), and opiates (N=6). The review includes a careful analysis of why cue exposure therapies in many cases fail, why they often succeed, and what specific factors determine their degree success. Conklin and Tiffany not only review the clinical and field studies with human subjects, but also cite the most current animal research on addiction extinguish to buttress their analysis. This is an academic paper, but clearly written and accessible to the intelligent layperson. For anyone struggling with addiction and willing to consider cue exposure therapy, I highly recommend reading this paper carefully to absorb its many insightful lessons.
The cue exposure treatment studies varied considerably in their design and execution. In about half of them (mainly the drug studies), the participants were abstinent during cue exposure. In one study (with alcohol dependence), a moderate drinking goal was encouraged by providing “priming” doses of alcohol, with the the prevention of drinking more than one drink, and this behavior was practiced both “inpatient” and as outpatient “homework”. Cue exposure varied from real “in vivo” cues to surrogate video, audio, and even “imaginal” cues just pictured in the mind. The frequency of cue exposure varied greatly — from a single cue exposure (e.g. smelling a glass of alcohol for 3 minutes) within a single session, to multiple, frequent exposures per session over 10 consecutive days of cue exposure sessions, to periodic exposure sessions spaced in time over weeks, with follow up over 6-12 months.
In their review article, Conklin and Tiffany identify 4 main “threats to success” (and corresponding success factors) that explain both why cue exposure did not work well in some cases and where it either was, or could be made, more effective. Before summarizing these success factors, I think it is important to note one key insight they highlight regarding recent learnings from animal research:
Rather than simply trying new things in an effort to discover the optimal parameters for use in cue-exposure addiction treatment, ideas for improving treatment can be directly informed by recent animal learning research focusing on extinguishing learned behavior…ideas about extinction have changed considerably since cue exposure was first introduced as a treatment for addiction. For many years, extinction training was believed to lead to a weakening of the initially condition CS-US association…However current concepts about extinction resemble more closely the original ideas of Pavlov (1927), who postulated that repeated unreinforced exposure to the CS does not break the original CS-US learning, but rather serves to mask it…Therefore, the conventional notion that extinction is unlearning has been replaced with the position that extinction is new learning, that is, during extinction, CS-US learning remains intact, but new associations develop to the original CS. (p. 159)
This is a crucial insight! The original addictive response to stimulating cues will never die by itself, merely by not reinforcing those stimuli. Rather, it is important to learn new behavioral responses to those old cues which come to “mask” or dominate the the old responses. Cue extinction is an active process, not a passive one!
Now let’s turn to the specific threats to the success of cue extinction which have been identified by Conklin and Tiffany:
- The renewal effect occurs when a behavior is successfully extinguished in one limited context or set of cues, but re-emerges in response to a different context or cues. This is a common problem in treatment, because the treatment context often differs in significant ways from the “real world”. Conklin and Tiffany give the example of a heroin addict who gets inpatient extinction treatment in a hospital room, but resumes shooting up at home–a different context, with different cues. The same conditioned stimulus (CS)– for example seeing or handling drug paraphernalia, or being stressed–can acquire a different “meaning” in the two different settings. Cues can be rich, subtle and varied: the action of lighting a cigarette with a match, handling of drug equipment, or the smell, the size and feel of surroundings, people, and the time of day. There are a number of important ways to deal with this problem. First, the extinction training should as much as possible occur in the “original conditioning context”, that is the real-world context in which the addiction was acquired and has been developed. Second, given the fact that most addictions are reinforced by a rich set of cues and multiple contexts, the extinction training should occur in several distinct contexts, and then re-tested in the original context. According to the authors, “Apparently, whereas conditioning generalizes readily, extinction is largely context-dependent”. (p. 160).
- Spontaneous recovery occurs merely with the passage of time, even when a behavior is initially extinguished successfully. The addiction can re-emerge by itself days, weeks, or months after being apparently terminated. Dealing with spontaneous recovery requires consideration of the “temporal spacing” of cue-exposures. Here, the authors cite a number of animal studies for guidance. In one such study, extinction occurred more rapidly and successfully when the cues were given as a series of short exposures over time instead of as a single “massed” presentation. Other studies found that extinction success was optimized by allowing longer intervals of time between exposure sessions, combined with more frequent in-session exposures. This was also reflected in the human studies. Based on this research, Conklin and Tiffany give the following guidelines:
- Within each session, the cue should be presented several times to ensure complete extinction of “responding”, defined as as subjective desire or objective physiological or behavioral response
- Within-session exposures should be separated by enough time to allow some recovery of responding between exposures
- Enough time should be allowed between sessions to allow for spontaneous recovery of responding, and therefore further extinction at each session
- The number of extinction sessions needed depends on the individual’s pattern of responding, which can vary considerably among individual subjects
- Reinstatement occurs after a conditioned stimulus (CS) has been extinguished, by presenting the unconditioned stimulus (US) alone. For those not familiar with this terminology (which is described in more detail on the Psychology page of this blog), the US is the immediate agent that produces the addictive “high”, e.g. the drug, tobacco, alcohol or food itself, whereas the CS is any cue which becomes associated with it, e.g. seeing or handling a bottle or cigarette, or visiting a bar or drug dealer. So in reinstatement, the former addict has learned not to respond to the environmental context and cues, but for one reason or another encounters the addictive substance in a new context, re-igniting the addiction anew and leading to potential relapse after even a single new exposure. Here, the research on prevention is very interesting. Relapse in such situations can apparently be prevented or quickly cut off by immediately exposing the lapsed addict to unreinforced exposure to the new context alone (without the US). So if your addiction to sugar or alcohol is re-ignited by inadvertently or unwittingly consuming a food that stealthily contains this offending substance, expose yourself to eating other foods (without the addictive substance) in the same place and with the same cues, on more than one occasion, and the relapse will be forestalled.
- Behavioral cue conditioning is one of the more subtle, but insidious threats to successful extinction. If the addiction is based upon classical conditioning (that is the addictive behavior is a direct “conditioned response” (CR) to one or more conditioned stimuli (CS), then deconditioning by extinction training has an excellent chance of success. However, in many cases of addiction the CS indirectly elicits behaviors that precede the direct addictive response, and these behaviors themselves act as secondary “discriminant stimuli” which provoke the addictive response independently of the CS. For example, for an alcoholic, the CS may be a bottle of booze. By the principles of classical conditioning, the appearance of the bottle can be extinguished as a cue for the urge to drink (the CR or conditioned response), by exposing the alcoholic to the bottle and not allowing drinking. However, in the normal context, the alcoholic engages in certain active routines or behaviors, such as pouring the alcohol into a glass, handling the glass, drinking from the glass, etc. These behaviors in themselves serve as independent cues, beyond appearance of the bottle itself, that stimulate the desire for the alcohol. So it is not just the sensory stimuli that need to be extinguished, it is also the behavioral cues. Overlooking this reality turns out to be a major flaw of many of the less successful treatments reviewed by Conklin & Tiffany. In these flawed treatments, the cue exposure sessions dealt with sensory cues alone. The authors found the best treatments involve extinction of active behaviors. For example, one study had smokers actually light cigarettes and take non-inhaled puffs. Another study had heroin addicts go through an actual cook-up procedure and handle all their paraphernalia, without allowing follow-through to actually administering the drug. While to an adherent of the “abstinence” approach such therapies may seem unduly risky, the science actually supports such realism as being the most effective way to immunize an addict against relapse.
There is some recent evidence from a study by researchers at McMaster University and the University of California at San Francisco that takes this approach even further. In cases where the goal is moderation and not abstinence, it is important the the cue exposure involve actually take small doses (e.g. one drink), while preventing any follow up drinks, to re-train the response. This is based on observations that addicts or alcoholics respond to a small dose as a cue that “more is coming”. Without this type of conditioning, there may be increased risk of relapse. Again, “you get what you train for”.
Conclusions. In short, cue exposure therapies will not work if they are confined to small number of artificial exposures within a single limited context, especially if it is significantly different from the context where the addictive behavior was “learned”. The exposure should be rich and varied, repeated both within a cue exposure session and at subsequent sessions while allowing an adequate time interval both between in-session exposures and between separate sessions to allow “responding” or partial re-emergence of the desire or craving. Cue exposure therapy should not involve mere passive exposure to sensory cues but should include a realistic “behavioral” component which is practiced without allowing the reinforcement itself to occur. Finally, it is important to keep in mind that extinction is not a matter of passively “unlearning” an old behavior by just not responding, but actively learning new substitute behaviors for responding to the original cues and contexts; adding a degree of “counter-conditioning” is useful here (see the discussion of counter-conditioning on the Psychology page of this blog.
What does this mean for you? Is there an addictive behavior or a bad habit you would like to overcome? Are you willing to try cue exposure therapy. If so, observe and think about the sensory and behavioral cues that precede your behavior and how you could design your own cue exposure sessions to help extinguish the behavior.
What does this mean for me? I stated at the beginning of this post that I would do something unusual. Rather than writing this post purely as a scientific report or as an “advice column” to others, I am going to put it to the test on myself. In the tradition of self-experimentation inspired by Seth Roberts, I am going to put my money where my mouth is and try it on myself. I have used cue extinction already as the basis for deconditioning myself from having a strong appetite for food (at certain times of day), for cutting back significantly on certain favorite desserts (such as ice cream), and for giving up caffeinated coffee (but still enjoying the occasional cup of decaf). However, I retain a certain fondness for alcohol. I’m not an alcoholic and and don’t believe I have a drinking problem, but I drink more than I would like to and find myself craving certain drinks before dinner almost nightly. My favorite drinks, in order, are: (#1) B&B cognac liquor on the rocks; (#2) Manhattan cocktail; (#3) beer; (#4) red wines, especially Pinot Noir. About a year ago, I cut back to a frequency of 1-2 drinks per week, but recently this has crept up to a nightly drink, and I find myself really looking forward to it after work. It is a real pleasure and stress reliever, and I don’t want to cut back, but I know I should.
So you’ll find a record of my experiment, starting today (Thursday, April 14), on my personal page on the Discussion Forum. At this point, my goal is not total abstinence, but cutting down to a maximum of 1-2 drinks on 1-2 nights per week. Wish me luck!