In their article "The Neural Mechanisms of Drug Addiction," Hyman, Malenka, and Nestler (2006) approach the long-standing problem of drug addiction from the point of view of understanding the addictive nature of drugs as being rooted in the ability of the addictive substances to hijack the brain circuitry normally involved in reward-related learning. Specifically, they believe that the cause for the compulsive nature of drug addiction (i.e., the fact that addicts continue self-administering the addictive substances despite full knowledge of the negative consequences of doing so) is to be found among the same areas of the brain as are responsible for associative learning and the formation of long-term memories. These areas of the brain include the ventral (or "underside of") and dorsal (or "upper side of") regions of the striatum, an area of the brain involved in the anticipation of reward, (Speert, 2012), which, as Hyman, Malenka, and Nestler (2006) go on to explain, is sensitive to electrochemical signals from dopamine neurons located in the midbrain.
Hyman et. al (2006) begin their article by introducing drug addiction, and especially the problem of relapse despite a patient's willingness to stop self-administering a particular drug, as a serious public health problem, the solution to which- the creation of improved treatment programs- can only come as a result of greater understanding of the specific neural processes involved in the addiction process itself, (Hyman, Malenka, and Nestlar, 2006).
While Hyman et al. (2006) note that a large amount of progress, especially involving the use of animal models, has already been made, (this has largely been possible because the drugs themselves can readily be identified as the cause of the addiction), they emphasize the importance- and the difficulty- of linking these findings obtained with animals (whose worlds can be carefully controlled by experimenters in the laboratory), to the experiences of human beings, whose worlds are far more complex and not subject to careful control and experimental manipulation. They argue, for example, that while much is known about the immediate and short-term effects of the binding of addictive drugs to specific sites in the brain, the question of the long-term effects of such binding and the changes in the brain they might engender merits additional research. Thus far, as they explain, it has been found that several distinct types of adaptation to long-term drug exposure occur, including those of homeostatic adaptation [this was discussed in a previous blog post in the context of opponent-process theory of human motivation, in terms of the gradual strengthening, over the course of repeated drug self-administrations, of a "b" process, which would come online to balance out the effects of a stimulus-dependent "a" process, (which, in this case, would represent the disruption of homeostasis caused by the self-administration of a psychoactive drug)- and how, over time, the "b-process," which would come to be initiated by cues related to the impending self-administration of a drug, would become stronger and come online sooner, resulting in a decreased overall effect of drug-taking and the phenomenon of the buildup of drug-tolerance (the need for an addict to take an ever-increasing amount of a drug of abuse in order to get the same effect as they did earlier, before the b-process became strengthened via repeated exposure to cues signaling the impending self-administration of a drug- and the disruption of homeostasis that would inevitably come as a result, (Domjan, 2009, p. 117))]- as well as, in the words of Hyman et al. (2006), "synapse-specific 'Hebbian' adaptations of the type thought to underlie specific long-term associative memory," (Hyman, Malenka, and Nestler, 2006, p. 566).
Hyman et al. (2006) begin their discussion by noting that while the amount and variety of natural stimuli that activate the reward circuitry of the brain is relatively large, the amount of substances which can "hijack" this circuitry, thereby mimicking the effects of natural rewards, is actually relatively small- limited to just "the psychostimulants (cocaine and amphetamine), the opiates, nicotine, ethyl alcohol, and the cannabinoids..." (Hyman, Malenka, and Nestler, 2006, p. 567). Interestingly, however, once they are exposed to these drugs, humans and other animals rapidly learn the cues that predict their availability, and (as is further discussed in previous blog posts), exposure to these conditioned stimuli initiates strong cravings for the drugs, and may lead to relapse among recovering addicts. Hyman et al. (2006), go on to describe this in terms of a conditioned place preference model- one in which, if rats or mice have been previously exposed to a pleasurable unconditioned stimulus (such as a drug) in a particular location, they will come to develop a preference for being in that location over other locations within the same enclosure or other general space (presumably due to the association which they have formed between the conditioned stimulus of being in that location and the unconditioned stimulus of receiving the reinforcing stimulus of the drug (Hyman et al., 2006). An interesting study of this was conducted by Akins (1998, Experiment 1). In that study, which involved male quail, the conditioned stimuli associated with being in a particular location within the cage, were conditioned to predict the availability of sexual reinforcement, in the form of the presence of a female quail. In that study, the male quail were placed in an enclosure which had two compartments, both of which looked visually appeared very different. After a test designed to establish the baseline preference of the male quail in regards to which of the two compartments they preferred to be in, the compartment which the quail preferred least was chosen to be the one for which preference would be conditioned (with the visual cues related to being in this compartment coming to be conditioned as the visual stimuli). The quail were then divided into two groups, an experimental group and a control group. The conditioning itself consisted of having one male quail from the experimental group at a time enter the chamber for which the male quail had originally demonstrated less preference, and remain there for a duration of five minutes, after which a sexually-receptive female quail was placed into the chamber with them, also for a duration of five minutes. Thus, for the male quail in the experimental group, the conditioned stimulus of the visual and other stimuli denoting the previously-less-preferred conditioning chamber were paired with the unconditioned stimulus of the appearance of a sexually-receptive female bird. For the quail in the control group, by contrast, the sexually receptive female quail was presented to them in a different location and a full 120 minutes before their subsequent exposure to the previously-less-preferred conditioning chamber and its related contextual stimuli. Thus, for the quail in the control group, the conditioned stimulus of the visual and other appearance of the previously-less-preferred experimental chamber and the presence of the sexually-receptive female quail were presented in what was essentially an explicitly unpaired fashion, preventing the conditioning of an association between these two cues among members of the control group. As was predicted, members of the experimental group of quail came to develop a preference for being in the previously-less-preferred chamber following the pairings of the visual and other cues representative of this chamber with the unconditioned stimulus reinforcer of the presence of the sexually-receptive female quail (Akins, 1998, Experiment 1).
Furthermore, Hyman et al. (2006) go on to mention that while repeated self-administrations of any drug will result in homeostatic adaptations within the circuitry of the regions of the brain stimulated by the drug, the specific means by which this happens and the resulting likelihood of the development of drug tolerance and addiction, and the ways in which such phenomena might become manifest vary markedly between different types of addictive substances, "... depending on the expression patterns of each drug's receptors and the signaling mechanisms engaged by the drug stimulation in relevant cells," (Hyman, Malenka, and Nestler, 2006, p. 567). Similarly, they note that substance dependence- which they characterize as the unmasking of the changes in the brain that have been caused by the self-administration of a particular drug which become evident as soon as regular drug self-administrations cease- can also vary greatly depending upon the drug taken and the specific neural circuitry involved. They go on to give the examples of how "... withdrawal from opiates or ethanol can produce serious physical symptoms, including flu-like symptoms and abdominal cramps (opiates), or hypertension, tremor, and seizures," (Hyman, Malenka, and Nestler, 2006, p. 568).
This variability in the withdrawal symptoms which come following an individuals' ceasing to self-administer a drug to which they have become accustomed, is consistent with Siegel's Conditioning Model of Drug Tolerance, according to which repeated self-administrations of a drug lead to the cues related to the impending administration of the drug coming to be conditioned to predict the physiological effects of the impending drug administration- which in turn leads to the earlier and increased activation of a "b process," which comes to attenuate the "high" or other pleasurable sensation experienced following an individual's self-administration of a drug, leading to the buildup of drug tolerance (Domjan 2009). As Domjan (2009) also explains, over the course of repeated conditioning trials, the stimuli related to the impending self-administration of a drug come to elicit a conditioned response . According to Pavlov's Stimulus Substitution model, which argues that, in the development of conditioned responding, the conditioned stimulus comes to operate in the exact same way as the unconditioned stimulus did previously, with the conditioned stimulus coming to activate the same neural connections as the unconditioned stimulus previously activated (Domjan 2009), cues related to the impending administration of a drug should then come to lead to the experience of the same physiological effects as the administration of the drug itself. While in many cases, the conditioned response to the self-administration of a drug is indeed just like the unconditioned response to the drug would be [an example of this is a study conducted by Ehrman, Robbins, Childress, and O'Brien (1992), in which two groups of participants- one, a group of former cocaine users, and another, a group of men with no history of using cocaine, were exposed to three experimental conditions: one in which they were exposed to cues related to the use of cocaine, one in which they were exposed to cues related to the use of heroin (none of the participants had any experience with using heroin), and another condition in which cues unrelated to the use of either drug were presented. Interestingly, in that study, the group of former cocaine users specifically experienced an increase in heart rate when exposed to the cues related to the cocaine-related stimuli (and not in response to the heroin-related or neutral stimuli)- a result exactly in line with Pavlov's Stimulus Substitution Model! In that particular case, the conditioned response to the conditioned stimuli of the cocaine-related cues came to evoke exactly the same response as the unconditioned response the former cocaine users would normally have to cocaine- that is, an increase in heart rate from baseline], this is not always the case; in fact, in many cases, the form of the conditioned response to environmental cues which have come to be associated with the self-administration of the drug is just like the form of the compensatory unconditioned response, rather than the primary unconditioned response. The case of the opioid drug heroin, for instance, is a great example of this. When heroin is administered, the drug itself is an Unconditional Stimulus (US) which elicits two distinct unconditioned responses- one of these is the primary unconditioned response, (this response has also been mentioned above in the context of the opponent process theory of motivation as being the "primary" or "a" process), which moves the system out of homeostasis, and another of which is the compensatory unconditioned response, (discussed above as being the "b process" in the opponent-process theory of motivation), which counteracts the effects of the primary unconditioned response and returns the system to homeostasis. In the case of heroin and many other drugs, the form of the conditioned response to drug-related cues is just like the form of the compensatory response; thus, while the primary unconditioned response to a self-administration of heroin would be a lower heart rate, lower blood pressure, and analgesia (decreased sensitivity to pain), the compensatory unconditional response is an increase in heart rate and blood pressure, and an unpleasant increase in sensitivity to pain. This stands in sharp contrast to what Pavlov's Stimulus Substitution Model would predict- but goes some distance in explaining why the withdrawal symptoms of many drugs (i.e., heroin, alcohol, etc.) appear to be almost "the opposite of" those of the primary unconditioned response to the self-administration of these drugs; these effects are seen because, in the case of these drugs, the conditioned response an individual might have to being exposed to cues related to the consumption of the drugs is similar to the compensatory unconditioned response, rather than the primary unconditioned response to them! The fact that this conditioned response may be similar to or different from the primary unconditioned response also might go some distance in explaining the extent of the variation in symptoms, as Hyman et al. (2006) describe them, which individuals withdrawing from different types of drugs might experience.
However, as Hyman et al. (2006) note, "Whereas avoidance of withdrawal likely contributes to ongoing drug use (especially with opiates, alcohol, and tobacco), it does not explain the most frustrating characteristic, from a clinical point of view, of addiction: the persistence of a relapse risk long after a person has ceased taking drugs," (Hyman, Malenka, and Nestlar, 2006, p. 569). Instead, Hyman et al. argue that "... the primary neural substrates of persistent compulsive drug use are not homeostatic adaptations leading to dependence and withdrawal, but rather long-term associative memory processes occurring in several neural circuits that receive input from midbrain dopamine neurons," (Hyman, Malenka, and Nestlar, 2006, p. 569). Such a model of drug dependence can also go further to explain why drug addiction can be so persistent and difficult to overcome; in the words of Hyman et al., (2006), "Long-term memories, unlike most homeostatic adaptations, can last for many years or even a lifetime," (Hyman, Malenka, and Nestlar, 2006, p.569). Given this, and the fact that, as they also mentioned, many instances of relapse are fueled by exposure to stimuli previously associated with drug cues, it becomes more easy to understand the possible reasons why individuals might suddenly resume a drug-taking habit, even years after initially attempting to (or perhaps even totally) "quitting."
Furthermore, while noting a possible role for emotional stress as well as the initiation of drug cravings following exposure to drug-related stimuli, Hyman et al. (2006) nonetheless go on to emphasize the role of brain pathways in drug addiction. Crucially, they emphasize that while drugs and other addictive substances do operate upon the same circuitry as natural rewards, the "hijacking," of natural reward circuitry in which addictive substances engage in is particularly detrimental, both because, "unlike natural rewards, drug rewards tend to become overvalued," (Hyman, Malenka, and Nestlar, 2006, p. 574), and because "... unlike natural rewards, addictive drugs do not serve any beneficial homeostatic or reproductive function, but instead often prove detrimental to health and functioning," (Hyman, Malenka, and Nestlar, 2006, p. 571).
Furthermore, they note that evidence suggests that, in the case of both natural rewards and those related to the self-administration of drugs, it is the increase in synaptic dopamine in the nucleus accumbens region of the brain. As Hyman, Malenka, and Nestlar (2006) go on to note, many of the neurons within the nucleus accumbens region of the brain have dendritic projections which allow for the binding of many molecules at once through synapses, with these neurons receiving signals from neurons that themselves release an opioid, made within the brain, which attaches to what is called a "mu receptor." Furthermore, more dopamine neurons in the ventral tegmental area of the midbrain also supply neurons within the nucleus accumbens region of the brain with additional dopamine. Psychoactive drugs can also influence the activity of neurons in the nucleus accumbens by influencing the release of both opioids and dopamine made within the brain, directly influencing the dopamine-sensitive neurons within the nucleus accumbens, or by having an impact upon the actions of the neurons which generate the inhibitory neurotransmitter GABA (and thereby regulate neural activity). Furthermore, neurons within the cortex can also have an influence upon the neurons within the nucleus accumbens, through their release of the excitatory neurotransmitter glutamate. Significantly, changes in how the post-synaptic (receiving) cell responds to the release of glutamate can trigger long-term changes in how a particular neural circuit operates- which in term can have important implications for learning processes (Hyman et al., 2006). In the specific case of the cortex and its release of glutamate onto the nucleus accumbens, the release of glutamate from the cortex is believed to provide the nucleus accumbens with crucial information regarding the engagement of particular sensory systems. To complete the picture, neurons in the cortex also release dopamine onto the neurons in the nucleus accumbens- with the latter neurotransmitter providing neurons within the nucleus accumbens with crucial information regarding the motivational state of the organism. When these two cues are taken together (and indeed, they often occur at the same time), the dopamine being released by the cortex can be seen as assigning a reward value of sorts to the level of engagement of various sensory systems (which can be gauged by the organism via noting the current level of release of glutamate neurotransmitter by the cortex) (Hyman et al. 2006). Crucially, what the animal seems to actually learn from this pairing of glutamate and dopamine release from the cortex onto the nucleus accumbens appears to be the difference between the level of reward it expected, and the level of reward it actually obtained, (Schultz, 2006). As Domjan (2009) notes, this learning-by-surprise parallels the way learning is hypothesized to occur in the Rescorla-Wagner equation, wherein, if a US is no longer surprising for any reason, including if it is already being perfectly predicted by a conditioned stimulus previously conditioned to asymptote at the time when an additional conditioned stimulus is added, then no additional learning will take place. In the case of dopamine release onto the ventral tegmental area of the nucleus accumbens, the same principle applies; if a forthcoming burst of dopamine from the cortex is already being accurately predicted by one stimulus, then another stimulus predicting the same phenomenon will fail to result in any additional learning taking place. In the case of drug abuse, psychoactive drugs have an unfair advantage in this process: since they can artificially ramp up dopamine release by the cortex- and this unexpected release of dopamine results in surprise, which stimulates the learning of a pairing between the sensory cues associated with the drugs of abuse and the dopamine release in the cortex- drugs can thus artificially acquire motivational value- which, tragically, makes the habit of using them all the harder to break.
Akins, C.K. (1998). Context excitation and modulation of conditioned sexual behavior. Animal Learning & Behavior, 26, 416-426.
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
Hyman, S. E., Malenka, R. C., & Nestlar, E. J. (2006). Neural mechanisms of addiction: The role of reward-related learning and memory. Annual Review of Neuroscience, 29, 565-598. doi: 10.1146/annurev.neuro.29.051605.113009
Schultz, W. (2006). Behavioral theories and the neurophysiology of reward. Annual Review of Psychology, 57, 87-115. doi: 10.1146/annurev.psych.56.091103.070229
Speert, D. (2012, February 02). Neuroeconomics: Money and the brain. Retrieved from http://www.brainfacts.org/In-Society/In-Society/Articles/2012/Neuroeconomics-Money-and-the-Brain