The Mind within the Brain

18
Addiction

Addiction is a disorder of decision-making. Defined this way, the question of addiction becomes one of identifying the different vulnerabilities or failure modes that have arisen within an individual’s decision-making systems. This hypothesis suggests that addiction treatments need to be tailored to the individual, based on the identification of those failure modes. Treatment processes are beginning to be tailored in just such a way.

Addiction is the classic disorder of decision-making—an addict continues taking drugs even in the face of massive negative consequences, and sometimes even in the case of a stated desire to stop.¹ How can an addict both state that he or she wants to stop taking drugs and yet continue taking them? As we’ve seen, human decision-making arises from a complex interaction of components. What we will see in this chapter is that this complex interacting set of decision-making systems, this decision-making machinery, can break down in many different ways. While each of these different vulnerabilities may produce subtle differences in how the addict acts (Does the addict crave the drug before taking it? Does the addict take only the one drug, or is any drug good enough? Is the addict chasing pleasure or avoiding dysphoria?), they all come down to the same problem—the addict continues the behavior, even though he or she should stop.

What is addiction?

Most modern definitions of addiction are based on the concept of maladaptive decisions—continuing to take drugs even in the face of severe negative consequences, whether they be illness, dysphoria, or longer-term negatives, such as loss of friends, family, job, etc.² But this concept can be taken to ludicrous extremes—do we really want to say that we are all addicted to breathing? If you take the oxygen away, we die; it’s hard to imagine a more negative consequence than that. More problematically, there are many cases where we celebrate an individual making a decision with severe negative consequences, such as a soldier defending his or her country, or a firefighter risking his or her life in a burning building. We celebrate Martin Luther King, who fought for civil rights, knowing full well that it was likely to end in his death. Sometimes, however, this view leads to interesting perspectives. The Russian poet Osip Mandelstam continued to write poetry even after he had been sent to the Soviet Gulag for it.³ The future Hall of Fame football quarterback Brett Favre continued to play football long after many suggested he stop, despite severe injuries and an aging body that was far less resilient to punishment than it had been.⁴ Should we say that Mandelstam was addicted to writing poetry or that Favre was addicted to football?

Other definitions of addiction have been based on the presence of a disconnect between stated desires and actions ⁵ (“I want to stop, but I can’t”). Of course, this definition implies that animals (who cannot state such a desire even if they feel it) cannot be defined as addicts. Given that animals reliably self-administer the same pharmacological substances that humans do, even in the face of negative consequences,⁶ this definition seems problematic. Another problem with defining addiction as a verbal disconnect is that many of us know addicts who do not state a desire to stop their addiction, who refuse to admit their “problem.” Many a high school clique has needed to intervene for a friend with a drug or alcohol problem, and there are many people alive today because their friends stepped in with just such an intervention.

Some economists have suggested that addiction should be defined by the elasticity of the decision. As discussed in Chapter 3, elasticity is defined by how much your decision to buy something changes in response to changes in price. As the economy worsens, people tend to go out to the movies or to dinner less often. These luxury items are highly elastic. On the other hand, people continue to buy food, even at high prices, so food purchases are highly inelastic. Addiction can then be defined as “irrational” inelasticity.⁷ As noted in Chapter 3, this allows economists to say that Americans are “addicted to oil”⁸—we continue to use it and to use it unwisely (driving individual cars rather than taking more efficient mass transit; depending on it for electricity and heating rather than using non–fossil-fuel systems or taking the time to ensure that our houses are energy-efficient), even in the face of severe negative consequences (major oil spills, high prices, and unnecessary wars).

It is important to note that drugs (like oil) are, in fact, economic objects;^A as the cost of drugs goes up, their use goes down, even among the most dedicated and physiologically dependent addicts.¹⁰ The number of people quitting smoking, for example, is directly proportional to the cost of cigarettes. Similarly, studies of changes in heroin or cocaine pricing (remember, the black market is still a market) find that increases in pricing lead to direct changes in the number of users and the amount of use. This elasticity is why one of the best ways to decrease the use of legal drugs (nicotine, alcohol, and, in the near future, marijuana) is to have “sin taxes,” which increase the price of the drug.

This elasticity is also one of the best arguments against the legalization of harder drugs (heroin, cocaine, methamphetamine). The argument for or against legalization is very complicated. On the one hand, the negative consequences of getting caught with drugs is a cost that decreases use. For example, studies of the amount of opium use in Taiwan during the Chinese crackdowns in the 1920s find a direct elastic relationship between the severity of punishment and initial use.¹¹ On the other hand, jailing users can interfere with treatment because it can reduce the positive options and positive interactions that are often closely related to recovery.¹² Jail time without treatment can throw users who have not yet committed violent criminal acts in with dangerous criminals who are, in fact, dangers to society. The other problem with criminalizing drug use is that it shifts drugs from an open market to a criminal market and tends to lead to violent business interactions, which have strong negative consequences for bystanders not actually involved in the criminal market.¹³ (See Prohibition, US 1920–1933; Columbia and the Medellin Cartel 1976–1993; and drug and gang violence in some American cities today.)

In part because of this complexity of defining addiction (and in an attempt to remove the stigma that had become attached to the word “addict”), psychiatrists in the DSM-IV and ICD-10 (regularly released books detailing the currently agreed-upon definitions in psychiatry) abandoned the word “addiction” and referred to drug users as “dependent” and overuse of drugs as “dependency.”¹⁴ I believe that this definition overemphasizes the pharmacological effects of drugs more than the interactions with decision-making. It also makes it hard to ask questions about “addiction” to nonpharmacological substances. I will continue to use the terms “addict” and “addiction” to refer to people making these maladaptive decisions to continue using drugs, while recognizing that the question of when one is an addict is more of a policy question than a scientific one. The scientific question is Why is the person continuing to pursue this drug? How can we change that? Whether we want to interfere, and how much we should be willing to push to interfere, is a personal question for affected individuals and family members and a policy question for society.

Addiction as a failure of decision-making

If we look at the decision-making system that we have explored so far in this book, what we have is a machinery of decision-making. Like any machine, there are multiple places it can break down. Looking at the system in detail, there are several dozen different ways that this machinery can make the wrong choice ¹⁵—the Pavlovian system can assign increased emotional motivation to a drug,¹⁶ the Deliberative system can misjudge the expected outcomes of a choice,¹⁷ the Procedural (habit) system can overvalue ¹⁸ or automate too quickly,¹⁹ the situation-recognition system can overseparate or overgeneralize situations,²⁰ etc. Each of these different failure modes or vulnerabilities could have an effect of driving the addict to continue making the addictive choice.²¹ Differences between these failure modes or vulnerabilities imply differences in how the addict makes those choices, and each one will likely require different treatments.

I do not mean to suggest that each addict’s behavior will be due to only one failure mode in his or her decision-making system. Biology tends to have lots of backup mechanisms. It may well be that full-blown addiction arises only when multiple systems break down. Drugs themselves also tend to create additional failure modes. For example, a massively pleasurable effect (the high of cocaine ²²) leads to increased motivation for cocaine (a “desire” for it ²³), as well as tolerance (requiring more cocaine to produce that high ²⁴), and an increased likelihood of taking actions leading to cocaine (the Procedural system, driving unthinking decision-making ²⁵). Alcohol leads to a reduction in anxiety (thus driving a learned need ²⁶), tolerance (thus requiring more and more alcohol to produce the same effect ²⁷), and reductions in the effectiveness of self-control and Deliberative systems.²⁸ Addressing only one vulnerability may not be enough to stop the addiction. On the other hand, sometimes one can find a key to allow the patient to find another path to prevent relapse.²⁹

To understand the concept of a failure mode or a vulnerability in the system, the simplest example is to look at opiate agonists. As we saw, there is good evidence that euphoria (pleasure) and dysphoria (displeasure) are instantiated in the brain by the flow of endogenous opioids (endorphin and dynorphin, see Chapter 4). This means that the feeling of pleasure is the flow of endorphins onto μ-opioid receptors. (This is a direct consequence of the rejection of the mind/brain duality that we have been discussing in the last few chapters.) Just as stimulating auditory nerves produces hearing sounds,³⁰ and waves of random activity across the visual cortex produce images,³¹ changing the flow of endorphins onto μ-opioid receptors changes euphoria signals. This means that a pharmacological chemical that tricked the μ-opioid receptors into thinking that they had seen endorphins would be experienced as euphoric pleasure. What are some chemicals that activate the μ-opioid receptors? They are the ones that gave us the words opioid and opiate, derived from the opium poppy (including such mainstays of addiction as morphine and heroin).³²

Addiction researchers will often talk of the drug as “hijacking” the neural system, but I think what is going on is something more subtle. We are taking an engineer’s view of the brain—there is a physical system and that physical system has ways it can break down. We evolved a machinery that includes a “pleasure” component as part of how it processes decisions (Chapter 4). This pleasure component is signaled by a chemical pathway. A foreign pharmacological substance is changing that chemical pathway. This takes us back to the “Car Talk” perspective that we started with in Chapter 2. To return to our original example at the start of the book, just as the thermostat in your house would react incorrectly if you held a match underneath it (it would think the house was hotter than it is), your thermostat would also react incorrectly if you passed a current over the wire that signals “I’m hot” to the thermostat. Heroin is changing that internal wire.

When Steve Jensen, Adam Johnson, and I looked at the decision-making machinery in this way in 2008,³³ we found many examples of these sorts of vulnerability points in the machinery; in fact, most of those failure points had already been found by other researchers. Many researchers had proposed theories of why addicts continue to take their drugs (for example, that their system gets out of balance and they need the drugs to return to balance;³⁴ that they are chasing euphoric properties;³⁵ that they learn to overvalue the drugs;³⁶ that the drugs become secondary reinforcers and draw attention away from other things;³⁷ that the drugs shift the balance between deliberative and habit systems;³⁸ etc.). What we found is that these theories were all stated in the form of “if you have this problem, you might be an addict.” What was interesting was that these theories were generally attacked by a researcher finding an addict without that problem. Logically, what the theories were saying was that a given problem would lead to addiction, not that all addicts had a given problem. Viewed in this way, addiction is a symptom that can arise from multiple underlying “diseases.” Our multiple-vulnerabilities hypothesis said that everyone was right—all of these problems could lead to addiction, because all of these problems were potential failure modes of the decision-making system. Figure 18.1 shows a list of some of the current known failure modes of the decision-making machinery.

It is important to note that addiction, by definition, is an interaction between nature and nurture. If one never takes cocaine, one never becomes addicted to cocaine and cocaine never takes over one’s life. Similarly, there are many different reactions to taking cocaine; there are many different reactions to smoking cigarettes.³⁹ Some people remain able to control their use, while others are lost to it after one experience with it, and others become lost to it only after extended experience. We do not yet know whether this is a function of the situations in which these users find themselves, or a function of the specific formulation of the drug itself, or the genetics of the user, or the learned background and life experiences of the user. (Most likely it is a complex combination of all of these.) What we do know is that drug addiction presents in many different ways, even among the population of people who are unable to control their drug use.

Craving: The difference between desire, withdrawal, and habit

One example of this difference is the now-clear clinical dissociation between craving and relapse.⁴⁰ For years, clinicians and research scientists have studied craving in addicts as a substitute for measuring what’s really important—relapse. In large part, they do this because craving is something that can be studied in the laboratory, in terms of a controlled, experimental paradigm. For example, in cue-induced craving, pictures of drug paraphernalia are shown to admitted drug addicts and to nonaddicts, and the addicts show more signs of craving (stress, sweats, nervous behaviors) and rate their feeling of craving higher on survey scales.⁴¹ However, it has been known for decades now that craving does not imply relapse. While craving and relapse are related (people who show a lot of craving are more likely to relapse, and people who relapse are more likely to show craving), they do not always co-occur. The problem is that while craving is a phenomenon of a moment, relapse is a long-term event that happens only (as an ethologist would say) “in the field.” Before we can actually measure this, we need to address what craving is, neurobiologically.

Craving is the inexorable desire for a thing—we’ve all felt it at one time or another. To quote the Rolling Stones, “you can’t always get what you want, but sometimes you get what you need.”⁴² The mechanism of craving has been studied extensively, particularly in terms of drug addicts being reminded of their addiction, for example by being shown pictures of drug paraphernalia. But craving is not such a simple thing. For example, it turns out to be dissociable from relapse in drug addiction. Even though someone might crave a cigarette or a drug hit, it is possible to resist that craving, and not relapse back into the addiction. Others may take the drug-taking action without ever craving it. As one patient with a gambling problem told a colleague of mine, “I don’t know what happened. I just found myself at the casino.” (Were they lying? Were they lying to themselves?) As we’ve seen, there are multiple decision-making systems driving our actions, only some of which are conscious. It is likely that if a person were to relapse through the Procedural system (Chapter 10), this relapse would not entail any craving beforehand (or perhaps even any prior recognition beforehand). They might well just find themselves “at the casino.”

Failure-point	*key systems*	clinical consequence
Moving away from homeostasis	motivation	withdrawal
(Koob and Le Moal, 2006; Koob and Volkow, 2010)
Changing allostatic set-points	motivation	physiological needs, craving
(Koob and Le Moal, 2006; Koob and Volkow, 2010)
Sensitization of motivation	motivation	incorrect action-selection, craving
(Robinson and Berridge, 2001, 2003)
Cue-outcome associations elicit pre-wired visceral actions	emotions and motivation	incorrect action-selection, craving
(Damasio, 1994; Bechara and Damasio, 2002; Bechara, 2005)
Escape from negative emotions	emotion	incorrect action-selection
(Koob, 2009)
Mimicking reward	multiple systems	incorrect action-selection, craving
(Volkow et al., 2002; Wise, 2005)
Errors in expected outcomes	deliberation	incorrect action-selection
(Goldman et al., 1999; Jones et al., 2001; Redish and Johnson, 2007)
Increased likelihood of retrieving a specific expected action-outcome path	deliberation	obsession
(Redish and Johnson, 2007)
Over-valuation of expected outcomes	deliberative	incorrect action-selection
(Redish et al., 2008)
Over-valuation of learned actions	habit	automated, robotic drug-use
(Di Chiara, 1999; Redish, 2004)
Timing errors	habit	preferences for unpredictable events
(Ross, 2008)
Over-fast discounting processes	deliberative, habit	impulsivity
(Bickel and Marsch, 2001; Bickel and Mueller, 2009)
Changes in learning rates	deliberative, habit	excess drug-related cue-associations
(Franken et al., 2005; Gutkin et al., 2006; Piray et al., 2010; Redish et al., 2008)
Selective inhibition of the deliberative system	system-selection	fast development of habit learning
(Bernheim and Rangel, 2004; Bechara, 2005; Bickel and Yi, 2008; Baumeister and Tierney, 2011; Bickel et al., 2012)
Selective excitation of the habit system	system-selection	fast development of habit learning
(Everitt and Robbins, 2005; Bickel and Yi, 2008)
Misclassification of situations: overcategorization	situation-recognition	illusion of control, hindsight bias
(Redish et al., 2007)
Misclassification of situations: overgeneralization	situation-recognition	perseveration in the face of losses
(Redish et al., 2007)

Figure 18.1 Some failure modes of the decision-making system that can lead to addiction. Obviously incomplete.

An important aspect of craving is that we crave things.⁴³ We do not simply crave pleasure; we do not simply crave. The word craving is transitive; it must include an object. We crave sex, we crave a cigarette, we crave a drink. Perhaps we can even crave love—but even then, what we are craving is a feeling, a moment, a romantic dinner, a walk on the beach. While that craving might be satisfied by something different from your original goal, there is no question that craving includes an identifiable target or goal. When someone craves something, there are extensive effects on the body itself.⁴⁴ Part of craving is that it increases arousal—the heart beats faster and harder, one’s attention increases, one gets jittery. These are Pavlovian actions that we saw in Chapter 8—untrained actions that occur in response to the delivery of the desired object that are now occurring in response to delivery of the expectation of that object.

Some people have explicitly identified craving with the Pavlovian system—something reminds us of the existence of an outcome (through a stimulus–outcome association), and we remember that the outcome was very desirable.⁴⁵ Just as we discussed when we introduced the multiple decision-making systems (Chapter 6), Pavlovian training has both action-selection effects (Chapter 8) and motivational effects (Chapter 13). It is not clear at this point whether the motivational effects of craving are identical to the physiological changes of craving or whether they just tend to co-occur.⁴⁶

Other people have identified craving with the Deliberative system, in that a Deliberative system entails a search for an outcome, finding paths that can reach to it.⁴⁷ This means that part of the Deliberative system must maintain a representation of the target so that it can be recognized when one finds it. Early studies of craving found that craving does not tend to occur for easily achieved goals; instead, it tends to occur for blocked goals. Consider an alcoholic who goes to the bar every day at a certain time and does not show craving until one day arriving to find the bar closed. Then the craving sets in.⁴⁸ This blocked path forces the alcoholic from a well-trained, regular habit-based (Procedural) decision-making system to the more flexible, more conscious planning (Deliberative) system.⁴⁹

When we discussed the motivation system (Chapter 13), we discussed a particular phenomenon called Pavlovian-to-instrumental transfer (PIT), in which Pavlovian associations reminded one of an outcome, which led to increased actions taken to achieve a specific goal. This suggests that craving may be an interaction between the motivational and Deliberative systems. Experiments directly aimed at PIT suggest a strong role for the nucleus accumbens shell, which is strongly involved in craving.⁵⁰ A relationship between PIT and craving would imply that craving is an interaction between the Pavlovian motivation (Chapter 13) and Deliberative (Chapter 9) decision-making components.

It is important to dissociate craving from the negative feelings associated with ceasing drug-taking. As quoted from a patient to Anna Rose Childress and her colleagues, “No, doc, craving is when you want it—want it so bad you can almost taste it … but you ain’t sick … sick is, well, sick.”⁵¹ Ceasing drug-taking produces dramatic withdrawal symptoms, both physical and mental. Physical withdrawal symptoms come because the drug affects more than just the central nervous system. Biological systems are fundamentally chemical systems, and the chemistry of the drug is going to have profound effects on the body. Of course, one of the points this book is about is that the mental changes are physical as well, and those chemical changes are going to have profound effects on both brain and mind.

Generally, withdrawal symptoms come from homeostatic shifts. The term “homeostasis” means maintenance of something at a given level, from the Greek words ομοιο (homeo-, meaning “same”) and στσις (stasis, meaning “standing still”). Much of biology is about maintaining the system at some point that will enable continued survival.⁵² These homeostatic levels are related to the intrinsic goals discussed in Chapter 13—when you need food, you are hungry; when you are too cold, you want to warm yourself up. The classic case of homeostasis is body temperature. Just like the thermostat that we saw at the very beginning of the book, your body contains temperature sensors (in the biological case, neurons throughout that project to a small nucleus of cells in the hypothalamus). If these sensors determine that the body is too hot, processes are initiated to cool the body down—capillaries near the skin surface open wider to provide more blood flow to the skin, allowing a faster transfer of heat from the internal core to the surface; pores open up, producing sweat, providing evaporative cooling. In extreme cases, humans will even open their mouths and pant. Similarly, if these sensors determine that the body is too cold, opposite processes are initiated to heat the body up, constricting capillaries, pulling the skin tighter (producing goose bumps), exercising muscles, etc.⁵³

These days, however, a number of scientists have recommended replacing the term with “allostasis,” from the Greek word αλλο- (“allo-,” meaning “other”).⁵⁴ In part, this redefinition is due to the implication that in homeostasis one maintains the parameter at a single target, while, in actuality, the target changes with one’s needs. (We used to hear that a normal human body temperature is 98.6 degrees Fahrenheit, but body temperature actually changes throughout the day and, of course, changes in fever.⁵⁵) In addition, biological systems learn to plan—if you are in a world where food is scarce, it’s better to overeat when food is available than to eat just until you’re full, because you don’t know if you’ll be starving later.⁵⁶ Part of the role of the Pavlovian motivation system (Chapter 13) is to guide that plan.

So what does this have to do with withdrawal symptoms? The key is that homeostatic systems react slowly. Imagine that your body is trying to maintain a certain oxygen flow through the bloodstream. If you move to a high altitude (say Katmandu in the Nepalese Himalayas), then getting enough oxygen to your tissues requires your body to put more resources into hemoglobin and oxygen transfer. This happens over the course of days or longer. Similarly, if you then move back to a lower altitude, the extra oxygen flow remains for a couple of weeks while your system readjusts. Athletes who require high oxygen flow (such as Olympic long-distance runners) often live or train at high altitudes so that their body will adjust and be better at delivering oxygen to their tissues when they run the actual race (although the efficacy of this is still controversial).⁵⁷

A cigarette smoker has been flooding his or her tissues with nicotine, which is an analog of a very common brain chemical called acetylcholine.⁵⁸ The music that I am listening to right now on my computer has three volume knobs in sequence. The media player has a volume, which changes how loud the music is that is sent to the computer, which has a system volume, which changes how loud the music is that is sent to the speakers, which have their own volume. If the media player volume increases, then I can get the same level of music out by decreasing the volume on the speakers (which is easier to reach than the internal system volume). But then if the media player volume suddenly decreases, the music will be too soft, and I won’t be able to hear it.

In response to the chemical flooding of nicotine, the brain has learned to turn the volume down on acetylcholine receptors. So when a person stops smoking, the body and brain react as if they weren’t getting enough acetylcholine—because the ex-smoker isn’t getting enough acetylcholine, because the volume on the acetylcholine receptors has been turned down too low. Such changes occur throughout the brain and peripheral systems. Withdrawal symptoms typically occur for a time after stopping drug use (days, weeks, months). These withdrawal symptoms are due to the body’s adjustment being out of alignment with the drug expectation.⁵⁹ Of course, as we discussed above, it is possible to change the body’s homeostatic (allostatic) target based on cues and situations. This means that cues can remind an addict of an expectation of drug, leading to a shift in the allostatic target, leading to what is effectively cue-induced withdrawal symptoms.⁶⁰

But craving can also occur in the absence of any explicit withdrawal symptoms.⁶¹ Craving is a mental process due to more permanent changes in learning and memory systems and can arise long after the withdrawal symptoms have gone away. Even though craving occurs through a physical process (everything in the brain is a physical process), understanding what craving is physically has been much more elusive and will likely depend on understanding how craving interacts with the multiple decision-making systems.

The key experiment would be to measure craving continuously as people live their lives, and then to compare those who report high craving levels and those who don’t and to compare those who relapsed and those who didn’t. This is, of course, a very hard experiment to do. There have been some attempts using cell-phones and PDAs to ask people to report craving at regular intervals.⁶² These experiments generally find that craving does often occur before relapse, but that sometimes craving occurs without relapse and sometimes relapse occurs without craving. Many readers may be asking why we don’t use retrospective experiments that find people who relapsed and ask them whether they showed craving beforehand. The problem comes back to the reconstruction of memory that we saw in the last few chapters. We don’t actually remember the episodes of our past; instead we reconstruct them from our memories of what was, what should have been, and what probably happened. In fact, careful experiments have found that people who relapse without craving often remember (incorrectly) that they showed craving before the relapse.⁶³

Craving can occur in an attempt to alleviate withdrawal symptoms.⁶⁴ Craving can occur because the emotional (Pavlovian, Chapter 8) system is expecting a strong reward (or needs to avoid a strong punishment),⁶⁵ or because the planning (Deliberative, Chapter 9) system recognizes a path to a strong reward.⁶⁶ It is not yet known whether these different drives will be dissociable experimentally or not, but they all entail mechanisms by which a person would express a strong desire for the drug.

Interestingly, all three of these mechanisms may lead to craving, but they don’t necessarily lead to relapse. The separation between craving and relapse comes back to the multiple decision-making systems that we have seen throughout this book. Craving is a function of the expectation of the outcome; you can’t crave something you don’t know you’re going to get. This means that craving won’t occur when actions are selected by the Procedural system (Chapter 10), which doesn’t include a representation of the outcome. Since the Deliberative system implies recognition of a path to a goal, but knowing that path does not mean one takes it (remember, the Deliberative system is flexible), one can have craving without relapse. Since both reflexes and the Pavlovian system can be overridden by self-control systems (Chapter 15), they don’t necessarily lead to relapse either. While craving and relapse are related, they are not identical.

Paths out of drug use

Although drug users, scientists, and society would very much like to have a single magic bullet that would “cure” a drug user of the predilection for drugs, the multiple-vulnerabilities theory of drug addiction suggests that there isn’t going to be a single path out of addiction.⁶⁷ That being said, current practice is to try multiple options, and some subset of these options often works for specific individuals.⁶⁸ My suspicion is that we need to see drug addiction not as the problem in itself, but rather as a symptom of an underlying problem or set of problems.^B

Although this book is explicitly not a self-help book, and I am not going to attempt to tell a drug user how to escape from his or her predicament,^C the multiple decision-making story that we have been telling throughout this book has some interesting implications for identifying potential paths out of the cycle of drug use.

The implication of this engineer’s view on addiction and the concept that there are multiple decision-making systems with multiple vulnerabilities provide two important perspectives on treatment that can be addressed. First, the effect of treatments should be to normalize vulnerabilities. It should be possible to predict treatment success from identification of outcomes of successful and unsuccessful treatment attempts. Second, different vulnerabilities will likely require different treatments. If we could identify which vulnerabilities an addict has fallen victim to, we could potentially provide the right treatment for those vulnerabilities.⁷³

Some evidence exists suggesting that these perspectives are correct—different addicts are better served by different treatments. For example, behavioral impulsivity measures predict success in smoking cessation. Neural structures and cognitive tests can be identified that predict treatment outcome success in cocaine addiction.⁷⁴

Some vulnerabilities will be simple to treat. A homeostatic imbalance can be alleviated by replacement with reduced-harm alternatives, such as nicotine replacement therapy for smoking or methadone replacement for heroin.⁷⁵ However, long-term relapse rates after ceasing these treatments are notoriously high.⁷⁶ This is likely a consequence of the fact that these therapies do not address the other vulnerabilities that an addict may have fallen victim to.⁷⁷

For example, it is important not to underestimate the negative component of drug addiction on emotion and affect. Because of the way that biological systems have limited internal resources, and because biological properties are always trying to balance themselves, a biological high (euphoria) will be followed by a crash (dysphoria).⁷⁸ If euphoria, for example, is the release of endorphins onto μ-opioid receptors, then heroin flooding the system will produce an overcompensating reaction like “ringing” in an oscillator. These negative effects can occur after even a single dose of morphine.⁷⁹ And, as we saw above in our discussion of homeostatic (allostatic) systems, these systems will learn to compensate for that overabundance of signal, leading to a negative assessment in contrast, when everything else just feels terrible. This is shown clearly in the movies Sid and Nancy and Trainspotting: the addicts’ senses have become so dulled to everything that drugs are the only thing left that can hold their interest for even a moment. Finding healthier activities that can substitute for drugs or alternative emotional factors that can capture one’s desires can be difficult but potentially useful. Sometimes the key is recognizing just how bad off one is and that alternatives do exist.⁸⁰

Some users have been able to break their addiction through attention to concrete alternatives, whether those alternatives are social (such as attending regular social interactions where one is praised for rejecting the addiction) or physical (such as earning money or points toward a new television set).⁸¹ There is good evidence now that concrete targets are easier to aim toward than abstract ones and that concrete targets draw more of one’s attention.⁸² When those concrete targets are immediately-available things one wants to reject, they are more difficult to reject. When those concrete targets are positive goals in the future, they are easier to attend to and easier to assign value to (thus making them easier to use as alternatives). I know a person who broke his smoking addiction by putting the money he usually spent on cigarettes in a jar every day. He was a jazz fan and set himself a goal of buying a new saxophone. When enough money had accumulated in the jar, he had quit his smoking habit and celebrated by buying himself the saxophone.

Some vulnerabilities can be exceptionally difficult to treat. As we saw in Chapter 10, stimulus–response decisions can be made without conscious attention or thought, and such habits can be very hard to break. We examined one potential solution in self-control, in that one can override those actions with frontal cortical self-control mechanisms (Chapter 15). However, those mechanisms require constant attention and are very tiring. Distraction, exhaustion, and stress can all make maintaining self-control very difficult.⁸³ Another alternative is to identify what cues lead to relapse and to avoid those cues. For example, an alcoholic might drive a different path home to avoid driving by the bar where he usually stops in for a drink.

This leads us to the concept of precommitment and bundling.⁸⁴ Imagine an alcoholic who knows that he will drink if he goes to a party that evening, but would prefer to stay sober when asked in the light of day. This alcoholic is, in a sense, two people—in the light of the day, sober and sane, and in the Pavlovian context of the party, a wild alcoholic. (This dichotomy of choice is, of course, the central theme of Robert Louis Stevenson’s Dr. Jekyll and Mr. Hyde.) These two temporally separated people are in conflict with each other.⁸⁵ If the sober and sane person in the light of the day (let us call him Dr. Jekyll) wishes to stay sober through the night, he only has to avoid the party. The problem is that in the evening, Dr. Jekyll will have changed his mind and will want to go to the party. One option is to find a way to commit himself to the nondrug option during the light of the day (for example, by making other plans for that evening). This is called precommitment because Dr. Jekyll has precommitted himself to an option so that the later option is not available.

Bundling is a similar process through which one changes the set of available outcomes by rethinking what the potential outcomes are.⁸⁶ For example, the concept “there is no such thing as one drink for an alcoholic” implies that one does not have the option to go to the bar and have “just one drink.”⁸⁷ Instead, the options are to not drink and be the boring Dr. Jekyll or to drink and become the devil Mr. Hyde. Being the cool Mr. Suave drinking one drink and charming the ladies is not an option for our protagonist Dr. Jekyll. Recognizing that there are only the two options (boring Dr. Jekyll and dangerous Mr. Hyde) leads our protagonist to a very different decision process (between the two options of Jekyll and Hyde rather than between the three options of Jekyll, Hyde, and trying to be Mr. Suave). Bundling and precommitment are two examples of ways that changing one’s perspective can change decision-making.

One interesting issue is that many (though unfortunately not all ^D) addicts age out of their addiction.⁸⁸ Whether this is due to changes in brain function with age (addiction often starts in adolescence, when prefrontal self-control functions are impaired ⁸⁹) or whether this is due to changes in life (such as gaining a family) is still unknown (and probably differs from person to person). Nevertheless, it is interesting that case studies of addicts who do age out usually talk of changes in perspective that lead to a rejection of the addictive choices.⁹⁰

There are too many treatment options available and their successes and failures are too complex to list here or to identify how each one interacts with the multiple decision-making systems. As noted above, different treatment options seem to work for different patients.⁹¹ Current treatments are now suggesting that patients go through a flowchart of options—if the patient has impulsivity problems, this can be treated with executive function therapy (which decreases impulsivity); if the patient has a homeostatic problem, this can be treated with pharmacological replacement.⁹² Other problems require other treatments.

Although there are some examples where treatments aimed at specific dysfunctions are identifiably successful,⁹³ a vulnerability-by-vulnerability treatment regimen is as yet unavailable. How to identify which failure mode is active (or which failure modes are active) in a given addict and how to treat those failure modes are a focus of active research throughout the neuroscientific and medical communities today.⁹⁴

Books and papers for further reading

• Gene Heyman (2009). Addiction: A Disorder of Choice. Cambridge, MA: Harvard.

• George F. Koob and Michel Le Moal (2006). Neurobiology of Addiction. Amsterdam: Elsevier Academic Press.