As noted above, very few studies have directly examined nicotine’s reinforcement-enhancing effects in humans, although other research has assessed various types of responding due to recent nicotine exposure that may relate to those effects. We describe some of that research here, to better define what is intended by the term “reinforcement enhancement” as addressed in the detailed studies comprising the main section of this review. (Those studies follow the description of methodology used in those studies to document such enhancement in humans.) For examples, acute nicotine via smoking increased subjective ratings of the attractiveness of pictorial faces (22) or of positive mood during a mood induction procedure in depression-prone smokers (23), and nicotine via patch increased correct responding on a challenging signal detection task reinforced by money in non-smokers (24). Further, nicotine via lozenge improved performance on a card-sorting task where speed of responding was reinforced by money, but it did so only in heavy and not light smokers (25), suggesting that nicotine-dependent participants were necessary. Comparable research found no effect of acute nicotine via “inhaler” on patterns of gambling behavior for monetary reinforcement (26). Because of their indirect or insensitive measures of behavioral reinforcement of non-drug rewards, these and related studies may be quite limited in what they can inform us about the existence of reinforcement-enhancing effects of acute nicotine intake in humans as demonstrated in the preclinical research. These studies also typically failed to assess responding for non-reward stimuli to verify specificity of nicotine’s effects on reinforced behavior per se rather than on non-specific responding (17).

Still other research has examined changes in similar types of responses as a function of smoking abstinence, instead of assessing acute effects of nicotine intake. This work generally proposes a disruption in incentive motivation or reward responsiveness due to nicotine abstinence in dependent smokers (27), rather than enhancement of reinforced behavior for non-drug rewards by acute nicotine administration. In one such study, overnight abstinence decreased self-reported pleasure from viewing positively valenced movie clips, relative to non-abstinent smokers and independent of withdrawal (28). On the other hand, smoking abstinence for one or more weeks in those attempting to quit permanently increased, rather than decreased, self-reported enjoyment of various “rewarding events” (29), opposite of that expected by the loss of nicotine’s reinforcement-enhancing effects. Thus, some of this research may relate more closely to withdrawal and anhedonic effects of recent abstinence (30, 31) or could reflect broader and more gradual improvements in overall well-being with successful abstinence (32), rather than loss of nicotine’s acute enhancement of responding reinforced by non-drug rewards.

Due to the reliable preclinical findings on nicotine’s reinforcement-enhancing effects (18, 19, 21), and the lack of prior human research directly examining these effects on reinforced behavior under carefully controlled conditions, we developed a procedure to assess the reinforcement-enhancing effects of acute nicotine on responding for non-drug rewards in humans. Our early piloting indicated a need for altering some methodological details (33) to finalize a procedure that was sensitive to changes in the reinforcing efficacy of some non-drug rewards after acute nicotine intake. Because most aspects of that procedure are common to the clinical studies to be described below, details of that procedure will be presented next, followed by results from the only published research examining factors moderating the presence or magnitude of nicotine’s reinforcement-enhancing effects in humans. (All are from our lab; we know of no other controlled studies specifically designed to assess these acute nicotine effects per se on reinforced responding for non-drug rewards in humans.) Conclusions of the human research, potential clinical implications of these reinforcement-enhancing effects for maintaining dependence and impeding success of tobacco cessation, and suggested future research directions are then addressed.

Our starting point was to develop and evaluate a testing procedure that matched those from rodent studies as closely as possible. We did so to minimize chances that failure to find reinforcement-enhancing effects of nicotine was due to procedural differences rather than lack of generalizability from non-human animals to humans. For example, to carefully control nicotine administration to participants via tobacco, their smoking exposure is determined by computerized instructions on the number, timing, and duration of all puffing behavior, confirmed by puff topography assessment after each administration. Second, our primary dependent measure is responding on a simple button-pressing task (to simulate lever pressing) that is reinforced by immediate delivery of brief rewards under a progressive-ratio (PR) schedule, which minimizes the odds of reward satiation. As the label implies, PR schedules require a progressively greater number of responses for each succeeding delivery of a reinforcing stimulus (i.e., a “unit” of reward) within a session, and the point at which such responding stops is taken as identifying the maximal reinforcing efficacy of that reward (17). Although occasionally used in human studies of reinforced responding for rewards (34), PR schedules are far more common in non-human animal studies (35, 36).

Also, to maximize statistical power (37) and assessment efficiency, testing involves a fully within-subjects design, in which drug condition (i.e., nicotine dose) is manipulated between sessions and the immediate reward condition varies across trials within sessions. Specifically, varying drug condition across three sessions allows us to assess the independent effects of acute nicotine per se (e.g., nicotine versus denicotinized cigarette) and of simply engaging in smoking behavior per se (denicotinized cigarette versus no smoking). Each condition is provided in counter-balanced order, and the order of reward condition is the same for a given participant across sessions so that only the session’s drug condition is different. (No order effects have been observed for either the drug condition across sessions or the reward condition within sessions.) All sessions start following overnight abstinence (confirmed by expired-air carbon monoxide of CO ≤ 10 ppm) so that subsequent testing of nicotine effects is based on controlled acute exposure due to study manipulations.

Reinforced responding for non-drug rewards using the PR schedule of reinforcement is assessed with a modified version of a simple computer task developed as a research tool for human studies of reinforcement schedules [“Applepicker” (38)]. It was previously used by us to assess acute reinforcement by access to smoking (i.e., nicotine’s primary reinforcing effects), food or money, or alcohol (39–41). To obtain rewards from using this task, subjects move a keyboard’s cursor around a video monitor to look for “apples” by pressing a button whenever the cursor lands on one of the small filled circles displayed in a grid, symbolizing an orchard of “trees.” The number of button presses (responses) required to find an apple is controlled by the computer’s programmed schedule of reinforcement. Participants themselves determine the duration of their responding, discontinuing only when they decide the response requirement for receiving the next unit of that reward exceeds what they are willing to complete. Total responding within a trial up to that point is used as the measure of that reward’s reinforcing efficacy; because participants will often continue responding after earning one unit of reward but stop prior to earning the next unit of reward, the “break point,” or final completed response requirement, alone may not fully capture the efficacy of that reward for a participant unless the additional responding is also counted. (This schedule was a PR30% in our first study, increased to PR50% in all subsequent studies, meaning each successive reward delivery requires 50% more responses on the task than did the preceding response requirement; e.g., 10 responses to earn reward the first time, then 15, 22, 33, 50, 75, and so on. The PR50% ensures a point of maximum efficacy for that reward is reached before the end of each 15 min trial.) A separate “no reward” trial, in which responding does not result in delivery of any stimulus, is routinely included in our testing sessions to allow assessment of non-specific responding and rule out nicotine’s general psychomotor stimulant effects on behavior. Moreover, responding on the “no reward” trial gauges whatever reward value is provided to the participant by simply engaging in the task [similar to the inactive lever in preclinical research (18)]. The monitor’s graphics are kept rudimentary to avoid the possibility that task responding might provide its own entertainment value in the absence of an available designated reward (i.e., serve independently as a reinforcing stimulus). Also, so that subjects do not simply continue responding on the task simply out of boredom, intentionally routine reading material (that would not be a competing reinforcer) is freely available until the trial’s end.

Finding an apple (i.e., completing a response requirement) earns an immediate reward delivery, or one unit of the designated reinforcer, as with onset of reward stimuli in the preclinical research (18). Each reward is available singly on separate test trials, which are presented in counter-balanced order within sessions across subjects. Thus far, the types of rewards enhanced by nicotine in preclinical studies appear qualitatively distinct (42) but may be characterized as stimuli that are “sensory” in nature [such as visual lights, auditory tones, olfactory, or gustatory tastes (19)]. Therefore, rewards in our human studies now usually include music (auditory), video (visual), and, for comparison, one that is not sensory (money) along with the no reward trial to gauge non-specific responding. Typically, units of reward for completing each reinforced response requirement are 30 s of preferred music or video (positive sensory reinforcement), $0.10 for money (positive non-sensory reinforcement), or no stimulus (no reward control). These music, video, and money positive reward units were selected based on piloting that demonstrated equal reinforcing efficacy between each under controlled conditions. Auditory stimuli are presented through headphones, video is shown in a separate viewing portion of the monitor, and money earned is displayed on a counter shown on the monitor.

Critically, the specific preferred music and video rewards are identified separately by participants during an introductory session by listening to their own supplied musical tracks, or by viewing identified video clips (via electronic means), and rating each on a 0–100 visual-analog “liking” scale, with those rated >75 considered “highly preferred.” This procedure of personalizing the music and video rewards ensures each will be clearly reinforcing for that individual, rather than assuming fixed music and video rewards will be comparably reinforcing for all participants (43).

We first used this procedure to examine the influence of nicotine (versus denicotinized) cigarettes on enhancing positive reinforcing effects of the sensory and non-sensory rewards of music and money, respectively, in 52 smokers (44). To further explore generalizability, we also assessed responding for the negative reinforcing effect of avoiding intermittent aversive white noise (for 30 s per reward unit) as well as for the no reward control of non-specific behavior. As noted above in the description of basic design procedures common across studies, participants were administered three different smoking conditions on three separate experimental sessions, each following overnight abstinence, to compare responses due to no smoking with those due to administration of 0.05 or 0.6 mg nicotine Quest brand cigarettes under blind conditions. Six puffs on either cigarette over 3 min preceded each trial on the two smoking sessions. Finally, our first study also confirmed that these effects do not require the presence of nicotine dependence or withdrawal, as in the preclinical research (18). To do so, we compared reinforced responding for non-drug reward due to nicotine in 25 dependent versus 27 non-dependent smokers, who averaged 14.3 versus 1.5 cigs/day, respectively.

Reinforced responding for music, but not money or avoidance of aversive noise (or no reward control), was significantly increased by the 0.6 versus 0.05 mg nicotine cigarette (see Figure 1A), confirming reinforcement-enhancing effects of nicotine intake per se on a positive sensory reward in humans. No differences in responding between smoking the 0.05 mg cigarette versus the no smoking condition indicated no effects of simple smoking behavior per se (without much nicotine) on any rewards. These findings due to nicotine were consistent between the dependent and non-dependent smokers, as no between-group differences in reinforced responding were found. Similarly, responding due to this smoking was unaffected by withdrawal relief, as elevated withdrawal at baseline in dependent smokers declined post-nicotine to that of non-dependent smokers, whose withdrawal score was low at baseline and remained low throughout sessions (44). Thus, there was no difference in reinforcement-enhancing effects of nicotine as a function of being a dependent or non-dependent smoker or due to nicotine’s actions in withdrawal relief. We believe this was the first direct demonstration of the reinforcement-enhancing effects of acute nicotine administration in humans.

The next step was to examine whether nicotine’s acute effects on enhancing reinforcement from music may vary due to the amount of acute smoke intake (i.e., nicotine dose), and the available reinforcer’s magnitude of efficacy (45, 46). Using a within-subjects design and procedures similar to our first study, 23 dependent smokers in study 2 completed 3 experimental sessions after overnight abstinence (47). The three sessions were virtually identical, differing only in the modest number of puffs from their preferred nicotine cigarette brand (i.e., unblinded) prior to testing: (1) eight puffs pretesting and then two puffs per trial, (2) two puffs per trial only, or (3) no smoking exposure at all. Each session involved separate trials of responding for reinforcement from high (rated >75 on 0–100 scale), moderate (rated 40–60), or low (rated 0–20) preference music rewards (or no reward control). As in study 1, rewards were personalized and identified separately by each participant during an introductory session, given individual differences in music preferences (43).

As expected based on their identified preferences, overall responding for reinforcement was influenced by the self-reported preference level of music reward. More importantly, responding for the high preference music trial, but not the moderate or low preference music or no reward trials, was increased after smoking eight puffs pretesting plus two per trial, compared to the two per trial only or no smoking sessions, which did not differ (Figure 1B). Again, withdrawal did not differ between the two smoking sessions, ruling out withdrawal relief due to nicotine as an explanation for the different responses for the high preference music reinforcement. These results confirmed that just over eight puffs from one cigarette after abstinence (as in a “lapse”) may be enough nicotine to obtain its reinforcement-enhancing effects, while very minimal smoking of just a few puffs is not. Findings of study 2 further showed the specificity of nicotine’s effects by its enhancement of responding for music reward that was high preference but not music of lesser preference (i.e., lower magnitudes of reinforcing efficacy) or for the no reward control (47).

Although the reinforcement-enhancing effect of nicotine on music reward in humans in study 1 was replicated in study 2, it was important to document its generalizability to other common sensory rewards in a smoker’s environment, in addition to music. Given preclinical research finding nicotine’s enhancement of responding for visual rewards (18), we chose to examine whether a visual reward would be enhanced by nicotine in humans similar to that observed with the auditory reward of music. Thus, in study 3 (48), we assessed acute effects of nicotine via smoking on responding for music and for video rewards, along with the same monetary reward or no reward control comparison conditions as in study 1. Preferred video rewards were identified for each individual in the same way as preferred music, as described previously. Also similar to study 1, a fully within-subjects design was used in study 3, in which 20 dependent smokers participated in three experimental sessions following overnight abstinence, differing only in the smoking conditions of six puffs on a 0.6 or 0.05 mg nicotine cigarette under blind conditions, or no smoking, prior to each task trial.

As hypothesized, reinforced responding for music and video rewards, but not for money (or no reward), was greater due to smoking the 0.6 versus 0.05 mg cigarette (Figure 1C), showing effects of nicotine per se (48). Lack of differences between the 0.05 mg cigarette and no smoking showed no effects of smoking behavior per se on reinforcement from music or video reward. Once again, effects were not influenced by withdrawal relief from either cigarette. Study 3’s results, extending nicotine’s reinforcement-enhancing effects from an auditory (music) to a visual (video) reward, were generally consistent with those from the preclinical research briefly noted earlier (19). They confirmed that acute nicotine intake per se enhances the reinforcing value of multiple sensory rewards (42).

An important factor needed to more closely link the human research to preclinical findings was to demonstrate that nicotine via a non-smoked manner of intake would also have reinforcement-enhancing effects. Studies 1 and 3 had manipulated nicotine exposure by administering cigarettes differing in nicotine delivery under blind conditions, so that all the non-nicotine factors accompanying nicotine intake would be controlled. To examine nicotine effects in the absence of tobacco smoke, we repeated the fully within-subjects design and procedures from study 3 (48), described above, but using electronic (e-) cigarettes differing in nicotine content (under blind conditions), rather than tobacco cigarettes. Thus, on three sessions after overnight abstinence, 28 dependent smokers in study 4 were administered an e-cigarette with vapor containing nicotine (36 mg/ml), one containing no nicotine, or no e-cigarette at all, prior to each trial in which they responded for the music or video sensory rewards, monetary reward, or no reward (49). Comparable to dosing in the smoking studies above, use of e-cigarettes was carefully controlled by precise instructions on timing and duration of 10 puffs over 5 min, a procedure shown to deliver rapid rises in blood nicotine (50).

Mostly similar to our prior studies of nicotine via tobacco smoking, reinforced responding for the video reward, but not for the other rewards (including music), was greater after the nicotine versus placebo e-cigarette (Figure 1D), confirming some reinforcement-enhancing effects of nicotine per se administered in a non-smoked formulation (49). Also as in study 3, no differences were seen due to behavioral effects of e-cigarette use per se without nicotine (i.e., responding due to the placebo e-cigarette in comparison with the no e-cigarette session). Replication of study 4 is needed to verify that music reward may not be enhanced by this amount and manner of nicotine intake from e-cigarette use despite very reliable music reward enhancement from nicotine via tobacco smoking (see studies 1–3 in Figures 1A–C), a result not easily explained. Aside from partly confirming that acute nicotine from a non-tobacco product has some reinforcement-enhancing effects in humans, as with tobacco cigarette smoking, these results could help explain the growing prevalence of nicotine e-cigarette use in the wider population (51).

Although much more research is needed, the clinical studies described here confirm the essential preclinical finding that nicotine enhances reinforcing effects of some non-drug rewards unrelated to nicotine. Study 1 indicated that nicotine per se acutely increases reinforced responding for music reward but not for money or the negative reinforcer of terminating aversive noise and that these effects occur with non-dependent as well as dependent smokers, as in the non-human research (18). Study 2 showed that nicotine from barely one cigarette acutely enhances responding reinforced by high preference music, but not by lesser preference music or from minimal cigarette smoking. Generalizability of reinforcement-enhancing effects of nicotine across types of sensory rewards, from music to video, in study 3 showed consistency with findings on reinforcing auditory and visual stimuli as rewards in preclinical studies. In study 4, some reinforcement-enhancing effects of nicotine from a non-tobacco, non-smoked method of administration (e-cigarette) was demonstrated for video reward, confirming effects in humans are not specific to nicotine via tobacco smoking. In all studies, the behavior of simply puffing on a tobacco or electronic cigarette alone did not alter reinforced responding, and withdrawal relief was unrelated to nicotine’s reinforcement-enhancing effects.

Given how commonly available many sensory rewards are in everyday life (52), including while smoking in environments where it is not restricted (53), an overlooked factor contributing to smoking’s persistence may be nicotine’s ability to enhance the reinforcing efficacy of these rewards. Recent U.S. population surveys indicate engaging in “leisure” activities containing such rewards may comprise about 5 h/day, or nearly one-third of waking hours (http://www.bls.gov/TUS/CHARTS/LEISURE.HTM). Therefore, an attempt to quit smoking also could lead to a loss of nicotine’s acute reinforcement-enhancing effects when quitters are engaged in non-drug activities providing sensory rewards, as clearly suggested by preclinical research on abrupt discontinuation of nicotine infusions (54).

To begin to understand the loss of nicotine’s reinforcement-enhancing effects that smokers may experience when initially attempting to quit smoking, we recently assessed responding reinforced by these sensory rewards during typical ad lib smoking compared to abstinence, in a two-session within-subjects design (55). Prior to the two sessions, 48 dependent smokers (non-treatment seeking) either abstained overnight (CO < 10 ppm) or smoked their own preferred brand in completely ad lib fashion (i.e., to capture typical nicotine satiation while smoking). During each session, they responded on the same task used in the above studies in three separate trials reinforced by the music or video sensory rewards, or for the no reward control. Relative to the ad lib smoking session, overnight abstinence significantly reduced reinforced responding for both sensory rewards but not for the no reward control condition (55), verifying sharp loss of these nicotine effects early in a quit attempt, similar to preclinical findings (54). As with our controlled studies of acute smoking after overnight abstinence, above, differences between sessions in withdrawal, which were substantial as expected, were not related to the differences between sessions in reinforced responding.

Consequently, quitting smoking may lessen overall reinforcement from these non-drug related rewards, separate and independent of withdrawal symptoms and other consequences of cessation. Although of uncertain impact without more direct clinical study of these changes, a decline in the rewarding effects of these activities could possibly contribute to a greater risk of lapse (and relapse) in an effort to restore this loss in reinforcement. Perhaps consistent, recent clinical research suggests early lapses after starting a quit attempt are common during leisure activities, including while concurrently enjoying “TV/music” (56). In short, aside from nicotine’s primary and secondary reinforcing effects, tobacco smoking may be particularly difficult to quit for long partly because acute nicotine makes engaging in so many of these typical daily activities more pleasurable. Therefore, a sudden lessening of this degree of pleasure from familiar rewarding activities after abstaining from smoking may help prompt a lapse, which then could foster complete relapse (57).

In addition to moderating craving and other withdrawal symptoms of abstinence, as well as blunting nicotine’s primary reinforcing effects, medications to aid smoking cessation could partly act by attenuating loss of nicotine’s reinforcement-enhancing effects after quitting tobacco use. No clinical research has formally addressed this notion, but some human studies are suggestive. For example, our findings above in study 4 showed potential for nicotine via e-cigarettes to prevent some of the loss of reinforcement-enhancing effects of nicotine via tobacco smoking after abstinence [although e-cigarette use has not been clearly demonstrated to aid smoking cessation (58)]. Other non-smoked nicotine products, especially the various U.S. FDA-approved nicotine replacement therapy (NRT) medications intended to help smokers quit (e.g., patch, lozenge, and gum, inhaler), may also enhance reinforced responding for sensory rewards. If confirmed in clinical research, these actions could further help explain why NRT reduces lapse risk and can prevent smoking lapses from turning into complete relapses (59).

Moreover, non-nicotine medications for cessation may also aid quitting partly by attenuating the drop in reinforcement from sensory rewards due to loss of nicotine intake after abstinence. Because of preclinical research showing that reinforced responding for rewarding stimuli was enhanced by bupropion (60), we conducted a small pilot study of this notion with bupropion in smokers recruited due to their high interest in making a permanent quit attempt. All were tested on three separate sessions, first after ad lib smoking and then during a crossover comparison of bupropion versus placebo, each after initiating a brief “practice” quit attempt period. In those meeting 24-h abstinence criteria during both medication conditions (CO < 5 ppm), reinforced responding for the high preference music reward decreased significantly, by almost half, when they quit while on placebo, while their responding after quitting on bupropion was similar to that during the ad lib smoking session (61).

Based on these modest findings, formal testing of NRT and of bupropion in attenuating loss of nicotine’s reinforcement-enhancing effects among quitting smokers may be obvious clinical research to conduct. Another may be to similarly test effects of varenicline, a partial agonist of nicotine that is FDA approved for cessation, on enhancement of reinforced responding. In rodents, moderate-dose varenicline enhances reinforcement from visual stimuli (62, 63), but other research indicates varenicline’s effects are less robust than those of nicotine, perhaps consistent with varenicline’s partial agonist actions (64).

Also needed are studies to more fully characterize nicotine’s reinforcement-enhancing effects in humans. These include determining other types of rewards that are enhanced by nicotine, potential individual differences in these enhancing effects by nicotine, and the possibility that nicotine may enhance a reward by attenuating habituation to its reinforcing efficacy, all of which will be addressed next. In addition, identifying what other drugs may have reinforcement-enhancing effects in humans could help determine pharmacological mechanisms contributing to these effects, as well as broadening the focus of enhanced reinforcement from non-drug rewards to other substance abuse problems. Preclinical research provides several directions to guide this human research.