ARU home page | ARU Profile | Experimental Methods | Biological Basis | Research Findings
University Courses | Opportunities | Research Reports

Paper presented to the FDA's Substance Abuse Advisory Committee meeting, 2 August 1994, Silver Springs, MD.

Is Nicotine Addictive? A Re-evaluation of the Data

Michael A. Bozarth, Ph.D.

Department of Psychology
State University of New York at Buffalo
Buffalo, New York 14260-4110


My name is Michael Bozarth. I am an Associate Professor of Psychology with the State University of New York at Buffalo. I am here today to express my views on drug addiction, its underlying biological basis, and the widely popularized notion that nicotine is an addictive substance. Although I have not previously expressed my position formally, I welcome this opportunity to "speak out" on a topic of serious concern for the scientific community--that topic concerns the hasty conclusions draw from an inadequate empirical database. I strongly believe the scientific community is responsible for presenting unadulterated 'facts' so the public and the government can make informed decisions. The lay public trusts scientists to adequately interpret their data and to remain unbiased by political or other pressures.

In this brief statement I will develop the argument that research has failed to substantiate the claim that nicotine is addictive. To the contrary, it is difficult to document even mildly rewarding effects from nicotine. Because of time limitations, I will restrict my presentation to a few key issues. The evidence presented here is exemplary of a much larger series of arguments that refute the main conclusions drawn from the 1988 Surgeon General's report. An exhaustive, point-by-point critique of that report is not possible with the limited time allowed for preparation. My main objective today is to broaden the debate by discussing key points that have not received adequate consideration.

By way of background, my main research interest is in the biological mechanisms of motivation and reward with specific application to drug addiction. Other interests involve motivational theory, biological psychiatry, and experimental methods used in psychopharmacology research. I consider myself a "bench" scientist, with daily involvement in the research activity of my laboratory. This "hands-on" approach to research provides a sharp perspective on the empirical data in an age when most senior investigators only 'manage' the research activities of their laboratories. In addition to conducting research, my university appointment also involves full-time teaching in my specialty areas (e.g., motivational theory, drug addiction, psychopharmacology).

The Nature of Addiction

The term addiction generally refers to the situation where drug procurement and administration appear to govern the individual's behavior, and where the substance seems to dominate the individual's motivational hierarchy. Two features that distinguish addiction from other behaviors are its extreme motivational strength and its motivational toxicity. Motivational strength refers to how hard the individual will work to obtain the substance, while motivational toxicity describes the substance's ability to disrupt the individual's normal motivations. This latter characteristic is particularly interesting because it may serve as a defining characteristic of addiction.

The individual is normally motivated to engage in a variety of behaviors. The relative importance of these rewards can be ranked to form a motivational hierarchy. Some motives will rank high, being extremely influential in the individual's behavior, while others will rank low, being relatively unimportant. The relative positions of motives on this hierarchy will vary from individual-to-individual as do what motives are even included on the list. Certain motivations are shared by all mammals (e.g., food, water, sexual behavior), while others appear unique to humans (e.g., career, television viewing).

Addictive substances have the ability to disrupt this motivational hierarchy. They do this in two ways. First, they can rapidly displace other motives in the individual's life, thrusting themselves to the top of the motivational hierarchy. Second, they can disrupt the ability of other, natural rewards to motivate behavior. The drug addict characteristically places substance use as their top priority, while losing interest in life's other rewards (e.g., food, sex). The intense desire to experience the effects of the addictive substance combined with the inability of natural rewards to engage behavior is subjectively experienced as a "loss of control." This perception, in a sense, is correct: the normal controls on the individual's life have lost their significance and behavior focuses on procurement and self-administration of the addictive substance.

Animal Models of Addiction

Several animal models of human substance addiction have been studied. These models can be divided into physiological effect models and reinforcement models. Physiological effect models include measures of physical dependence and withdrawal reactions and also some measures of neurochemical activity. Identified physiological effects of addictive substances are presumed correlates of their addictive properties, and research examines the characteristics of these effects. This approach, however, does not directly validate the model; many physiological effects of a substance are unrelated to its addictive properties.

Reinforcement models, on the other hand, directly study the reinforcing effects of a substance. This experimental method offers the closest model of human substance addiction. Although there are several methods of studying substance reinforcement, the intravenous self-administration method enjoys the most widespread acceptance among behavioral scientists. Most substances that are addictive in humans are readily self-administered by laboratory animals. This has led to the use of intravenous self-administration as a screening technique for potential addiction liability. Substances that are addictive are potent reinforcers in this behavioral paradigm, and compounds supporting self-administration may have significant addiction potential. However, many substances that are not considered addictive also reinforce behavior. It is important to keep in mind that evidence that a substance is reinforcing does not constitute proof that it is addictive. Additional tests must be conducted to form this conclusion.

Intravenous Self-Administration: Prototypic Addictive Substances vs. Nicotine

There are a number of characteristics shared by prototypic addictive substances such as cocaine and heroin. These substances have been tested extensively in intravenous self-administration, and the general properties of their self-administration are easily replicable across different laboratories. By contrast, these consistent findings from self-administration experiments with the prototypic addictive substances are not present with nicotine. I would like to briefly summarize some of the characteristics of prototypic addictive drug self-administration and compare these effects with those reported for nicotine self-administration.

1. Addictive substances such as cocaine and heroin characteristically support rapid acquisition of intravenous self-administration in a large percentage of the experimental animals tested. By contrast, few laboratories have reported successful nicotine self-administration. There have been numerous published failures with nicotine, and yet the scientific literature is always strongly biased in favor of positive findings.

2. Special testing conditions are not necessary to demonstrate the potent reinforcing effects of true addictive substances such as cocaine and heroin. By contrast, special testing conditions are necessary to demonstrate nicotine self-administration.

A. Food deprivation or restricted feeding schedules are frequently used.

B. Many studies with monkeys used stressed subjects:

(1) most studies house the monkeys in social isolation, and
(2) restraint chairs are frequently used to immobilize the subjects.
C. Most studies reporting nicotine self-administration have not used experimentally naive subjects: (1) some subjects have previous training with prototypic addictive drugs, while
(2) other subjects have previous training with a food reinforcer.
3. Self-administration behavior of prototypic addictive drugs is maintained with a variety of reinforcement schedules.
  A. Continuous reinforcement schedules (CRF) are always effective with such addictive drugs; and

B. Partial reinforcement schedules (e.g., FR, FI) have varying degrees of efficacy in maintaining behavior.

C. Again, nicotine is different. Positive findings with nicotine are reported with only a few reinforcement schedules.

(1) Indeed, CRF schedules have been consistently reported not to work with nicotine. (Yet this reinforcement schedule should be the most effective in establishing behavior.)
(2) And some of the reinforcement schedules that have been reported to work with nicotine are also effective in maintaining responding with painful electric shock.
4. Response patterns for a true addictive substance such as cocaine or heroin are predictable from its pharmacological activity.
  A. A wide range of substance doses is usually effective with such drugs. With nicotine, only a very narrow dose range is effective.

B. With addictive substances, behavioral compensation occurs following changes in unit dose (i.e., the amount of drug given with each injection). No such apparent regulation of intake has generally been seen with nicotine.

C. For prototypic addictive substances, specific receptor antagonists produce well-documented effects. These include increased drug intake following low antagonist doses and extinction-like responding following high antagonist doses.

D. By contrast, the effects of specific receptor-antagonist pretreatments in reported nicotine self-administration are anomalous showing a failure to observe:

(1) increased nicotine intake with low antagonist doses or
(2) extinction-like response patterns with higher antagonist doses.
5. A number of experimental control procedures have been used to insure the validity of the main conclusions from studies of prototypic addictive drugs such as cocaine and heroin; these procedures are designed to show that the substance is serving as a reinforcer. By contrast, control procedures are generally inadequate to show that responding is due to nicotine reinforcement.
  A. Most studies neglect the effect of mild stimulants on low-rate operant behavior; and

B. The two-lever choice test does not eliminate treatment x stimulus-presentation interactions--yoked control procedures need to be tested with nicotine to eliminate this possible experimental artifact.

6. Finally, intravenous self-administration of prototypic addictive substances is uniformly demonstrable across a high proportion of the experimental subjects tested, a wide range of mammalian species (e.g., rodents, monkeys, humans), and a number of different laboratories and scientific investigators. By contrast, positive findings with nicotine are reported from only a few laboratories.
  A. Nicotine self-administration appears to be very sensitive to minor procedural differences; and

B. The findings are somewhat controversial among drug reinforcement specialists.

Conclusions from the Nicotine Self-Administration Literature

The interpretation of the data is largely biased by the prior assumption that nicotine is addictive. If that assumption is suspended, however, it is clear the animal studies are inconclusive or even argue against the "nicotine addiction" hypothesis.

1. Nicotine is, at best, an equivocal reinforcer in intravenous self-administration tests. Few investigators have obtained positive findings under any testing conditions. If we were to accept the reports of nicotine self-administration as valid, then:
  A. nicotine is a reinforcer only under very limited testing conditions, and

B. nicotine reinforcement appears much less potent than reinforcement from natural rewards (e.g., food, water) which is considerably less potent than reinforcement from prototypic addictive drugs.

2. Nicotine clearly does not have the profile shared by prototypic addictive drugs.

3. For a drug (or any substance) to be addictive, it must serve as a reinforcer. But simply serving as a reinforcer does not constitute addiction unless we wish to classify all behavior as addictive. Motivational strength must be considered when evaluating the addiction liability of a substance. This will be reflected by four variables: the speed of acquiring the response, the resistance to extinction, the work output the subject will generate to obtain the reward, and the amount of aversive stimulation the subject will tolerate to obtain the reward. These variables must be considered if the intravenous self-administration of a substance by laboratory animals is to be used to decide addiction liability.

Nicotine & Brain Reward Mechanisms

The brain has specialized pathways that mediate motivation and reward. One of the systems studied most extensively is the mesolimbic dopamine system. This "reward" pathway appears to be involved in the reinforcing effects of a number of distinctively different rewards, including natural rewards such as food, sex, and maternal behavior. This pathway may also play a critical role in the genesis of drug addiction. Prototypic addictive drugs (e.g., cocaine, heroin) activate this pathway, as do other nonaddictive substances. The fact that two distinctively different pharmacological drug classes (i.e., psychomotor stimulants and opiates) may derive a major part of their reinforcing effects by activating a common reward substrate suggests a biological mechanism for a unifying theory of addiction. It is particularly enticing to speculate that other drugs with a clear addiction liability (e.g., barbiturates, ethanol) may also derive a major part of their reinforcing effects by activating this substrate. As with any theoretical model, the more behaviors the model can explain the more comprehensive the model. This has prompted speculation and study of the effects of other rewards on mesolimbic dopamine activity.

The fact that a substance can activate brain reward mechanisms does not suggest that the substance is addictive. Natural motivators also exert their influence on behavior by activating brain reward mechanisms. The pleasure experienced from cigarette smoking may involve specific brain pathways, but this is not fundamentally different than activation of these same pathways through sensory stimulation such as sweet taste or sexual gratification. It is simply a question of which neural element initiates the reward signal: peripheral sensory receptors or central neural components. The control addictive drugs exert over behavior is distinctively different from the modest effect produced by "life's little pleasures."

The exact role of brain reward systems in drug addiction is unclear. It is probably the case that activation of reward processes is necessary but not sufficient to produce an addiction. Reinforcement theory can be used to describe the relationship between the behavior and the drug presentation, but the mechanism appears more complex than simple activation of brain reward systems.

The "Voluntary" Nature of Cigarette Smoking

The issue of whether to regulate nicotine seems to focus on whether cigarette smoking is considered a voluntary act or whether it is the consequence of an addiction to nicotine and therefore not under the smoker's control. For a behavioral scientist, it is very difficult to reconcile the notion of "voluntary" behavior with the doctrine of empirical determinism. Cause-and-effect figures predominantly in experimental psychology. Behavior is viewed as the inevitable consequence of antecedent conditions, and the notion of free will is usually excluded from scientific consideration. However, upon pondering the question further it is possible to develop an operational definition of "voluntary" that satisfies the general sense conveyed by the layman's use of this term.

Voluntary behavior is considered behavior of choice. In more scientific terms, the behavior might be considered under the normal control of multiple factors and competing response tendencies. A number of motivations normally compete for the individual's attention. The desire to eat, to sleep, and to have sex may all motivate the individual at the same time. Which behavior predominates is the result of a complex interaction among the relative strengths of the motives, prevailing stimuli, and current physiological conditions. The normal interplay among various motives with no one motivation dominating the individual's behavior gives rise to a sense of self-control.

In the case of addiction, behavior is described as involuntary (i.e., loss of self-control). The motivational properties of the drug override the normal influence of natural rewards. This feature is more formally termed motivational toxicity. This is probably the result of the intense motivational strength of the drug reward and the diminished rewarding impact of other reinforcers.

The cigarette smoker experiences no such blunting of life's other rewards and pleasures. Indeed, smoking is frequently reported to enhance other pleasures. There is no extreme focusing of motivational priorities nor any disruption of the ability of other rewards to engage behavior. In short, there is no motivational toxicity displayed by cigarette smokers. Furthermore, behavior is still under the control of other motives and hence smoking must be considered "voluntary" behavior.

Challenges to the "Nicotine Addiction" Hypothesis

There are a number of challenges to the "nicotine addiction" hypothesis. Some of them concern the empirical studies and others are conceptual. Some of the scientific data have been accepted as fact without seriously entertaining alternative explanations. Hasty conclusions have been drawn from these data, frequently without considering elementary principles of experimental psychology and behavioral neurobiology.

Nicotine has only equivocal reinforcing effects in laboratory animals. These effects generally have not been replicated across laboratories, and the special testing conditions apparently necessary to demonstrate nicotine reinforcement have not been identified. Even if nicotine can be shown to reliably serve as a reinforcer, reinforcement from a substance is not equivalent to addiction. Otherwise, all behavior must be labeled addictive.

Reinforcement appears necessary to establish addiction, but reinforcement is not sufficient to produce an addiction. Most reinforcers are not "addictive." Drug addiction probably involves mechanisms other than simple reinforcement. Newer theories of addiction focus on neuroadaptive changes in brain reward mechanisms, and prototypic addictive drugs (e.g., cocaine, heroin) have been reported to produce changes in these brain systems.

Nicotine administration does not produce the primary characteristics of addiction. Specifically, neither extreme motivational potency nor motivational toxicity are associated with tobacco use. These effects are the hallmarks of addiction, and these effects are not typically associated with tobacco use. Nicotine does not propel itself to the top of the individual's motivations nor does it impair the ability of natural rewards to motivate behavior.

There is considerable confusion among the nonspecialists regarding the significance of activating brain reward pathways. Natural rewards activate these pathways and this process is part of the normal generation of behavior. Activation of brain reward mechanisms is not equivalent to addiction. Nicotine may activate brain reward mechanisms (e.g., produce a mildly pleasurable sensation), but so do all rewards!

For all these reasons, the data do not support the assertion that nicotine is addictive. Considerably more research needs to address the complex motivations for smoking behavior.

© 1994/1998 Addiction Research Unit/University at Buffalo

ARU home page | ARU Profile | Experimental Methods | Biological Basis | Research Findings
University Courses | Opportunities | Research Reports