|Home | About | Journals | Submit | Contact Us | Français|
Craving has long been considered central to addiction, yet it remains unclear just how it contributes to drug use or relapse. Improved understanding of its role in addiction requires a more fine-grained examination of craving, including the context in which it occurs. This study used a novel set of smoking-related and unrelated stimuli to investigate responses of 227 nicotine-deprived smokers under conditions that manipulated perceived smoking opportunity and motivation to quit smoking. Specifically, we contrasted features of preconsumptive states when active smokers anticipate smoking soon to states when smokers do not have this expectation. Results indicated that under certain conditions (smokers not interested in quitting who expected to be able to smoke soon), exposure to smoking cues becomes attractive. Specifically, these smokers found smoking-related images to be more pleasant than did smokers who were either motivated to quit smoking within the next month or who were informed that they would not be permitted to smoke during the experiment. More broadly, the study raises the possibility that the loss of routinely experienced pleasant cravings upon quitting smoking may yield challenges to staying quit, and requires greater clinical attention.
The topic of craving has interested addiction researchers for decades (e.g., Jellinek, 1955; Wikler, 1948). Clinical studies reveal a link between cravings and subsequent lapses (Sayette, in press; Tiffany & Wray, 2012). Although interventions designed to reduce craving have shown promise (e.g., Monti et al., 1993; O’Brien, Childress, McLellan, & Ehrman, 1990), clearly research that better accounts for contextual factors related to craving is needed both to improve outcome (Conklin & Tiffany, 2002; Heckman, Kovacs, Marquinez, Meltzer, & Tsambarlis, 2013) and to enhance conceptualization of craving (Wilson, Sayette, & Fiez, 2012).
A variety of learning-based models consider the relation between craving and addiction (see Niaura et al., 1988; Tiffany 1990). Negative reinforcement theories view craving as an unpleasant state, the alleviation of which motivates drug use (Siegel, 1983; Solomon & Corbitt, 1973; Wikler, 1973). Alternatively, appetitive motivational theories posit that drug cues come to induce a motivational state that approximates the actual effects of the drug (Stewart, de Wit, & Eikelboom, 1984). From this perspective, drug cues provide an incentive, experienced as craving, to receive more of the drug (see also Bindra, 1974). Although positive affect may trigger this drive, it is notable that the drive itself to consume need not be experienced as pleasant. As Stewart et al. (1984) state, “a conditioned reinforcing stimulus acts to maintain behavior not by virtue of satisfying some drive or need but rather by acting as a persistent goad to response generation.” (p. 263).
Accordingly, with few exceptions [e.g., Baker, Morse, and Sherman’s (1987) dual affect model of craving] cravings are considered either to be aversive states that must be alleviated (by using the substance) or merely goads to using (Stewart et al., 1984). Less investigated are the ways in which anticipation of drug use (separate from actual consumption) by continuing users not currently trying to quit, might contribute to the perpetuation of a drug habit. We tested two key contextual factors (perceived opportunity to smoke soon1 and current motivation to quit) thought to affect smoking anticipation. Specifically, we assessed how these two factors affect smokers’ experience of smoking-related (viewing images of smoking cues, holding a cigarette) and smoking-unrelated (listening to a comedy clip) stimuli.
The idea that drug availability influences the nature of craving is not new (Baker et al., 1987; Childress, McLellan, & O’Brien, 1986). Consistent with this premise, Juliano and Brandon (1998) and Droungas, Ehrman, Childress, and O’Brien (1995) found that smokers reported stronger urges when they expected than when they did not expect to smoke. In a review, Wertz and Sayette (2001a) observed a strong effect of perceived drug use opportunity, with those expecting to use reporting much stronger urges. Several subsequent studies across a range of measurement domains generally have reinforced the proposition that expecting to smoke soon increases urge ratings (e.g., Carter & Tiffany, 2001; Dols, van den Hout, Kindt, & Willems, 2002; Field & Duka, 2001; Thewissen, van den Hout, Havermans, & Jansen, 2005; Thewissen, van der Meijden, Havermans, van den Hout, & Jansen, 2008) and improves affect as assessed by self-report (Carter & Tiffany, 2001), electrophysiology (Zinser, Fiore, Davidson, & Baker, 1999), and affect-related facial expressions (Sayette & Hufford, 1995; Sayette, Wertz, Martin, Perrott, & Hobel, 2003). These data suggest that many cravings can be rewarding (pleasurable) anticipatory experiences, and that our usual ‘set’ that craving must be a frustrating moment of torment derives only from what we know about relapse, not ad lib smoking. Further, expecting to smoke soon also appears to affect the manner in which cues are processed. Research indicates, for example, that expecting to smoke soon heightens attentional bias toward smoking stimuli (Wertz & Sayette, 2001b), and alters neural responding to both smoking cues and to nonsmoking reward (Wilson, Sayette, Delgado, & Fiez, 2005; 2008).
Research indicates that individuals motivated to seek treatment experience different levels and types of cue-elicited cravings than do “continuing” drug users. In abstinence seeking situations, individuals do not intend to use the substance because they are motivated to quit. In abstinence avoidance situations, continuing smokers are not trying to quit, but their typical drug use routine is interrupted (e.g., no more cigarettes) and the aim is to regain access to and use the drug (Tiffany, 1990). During drug cue exposure, studies testing continuing users observed urges almost twice the level of those found when testing participants in treatment (see Wertz & Sayette, 2001a). Data from brain imaging studies also reveal different patterns of neural activation following cue exposure among these two types of users (see Wilson, Sayette, & Fiez, 2004).
The forgoing suggests that a particular combination of perceived smoking opportunity and (weak) motivation to seek treatment may elicit a type of craving associated with what Bradford, Curtin, and Piper (2015) refer to as anticipation of imminent use. We argue that this situation may arise when abstinent-avoidant (continuing) smokers who have not recently smoked (and thus are not already satisfied) are faced with the prospect of smoking very soon. That is, the experience of anticipation itself may become satisfying. Baker, Brandon, and Chassin (2004a) go further, suggesting that “there is more ‘reward’ in the anticipation of smoking than in smoking per se” (p. 477). Consistent with this position is the observation in monkeys that more dopamine is released during anticipation than during actual food consumption (Schultz & Romo, 1990; see also Volkow et al., 2006). In humans, a recent neuroimaging study examined responses among smokers varying in their motivation to quit and their perceived opportunity to smoke soon using the same four types of groups as in the present study (Wilson et al., 2012). Across multiple brain regions, connectivity analyses pointed to distinct activation patterns among continuing smokers anticipating imminent use, such that “greater engagement of the rostral prefrontal cortex was associated with greater activation of areas of the brain implicated in reward-related processing and craving…, perhaps reflecting the use of cognitive control to support positive anticipatory processing” (Wilson et al., 2012, p. 208). When smoking is imminent, continuing smokers may be most comfortable with, and most willing to “indulge” their cravings (Sayette, 2004). From an economic perspective, Loewenstein (1987) describes savoring as the “positive utility derived from anticipation of future consumption” (p. 667). Children who hoard their stash of Halloween candy, for example, may prefer savoring their candy to actual consumption. Accordingly, if smokers learn (perhaps unconsciously) to deliberately expose themselves to drug-related cues when actively using, then upon quitting some may inadvertently persist in exposing themselves to smoking cues that had triggered these positive anticipatory states (Baker, Japuntich, Hogle, McCarthy, & Curtin, 2006; Sayette, in press).
The prospect that anticipation of imminent use may offer key insights regarding smoking motivation is highlighted in a recent study. Bradford et al. (2015) observed that anticipation of imminent use among smokers not currently trying to quit served to relieve the stress associated with receipt of an electric shock. As these authors emphasize, preconsumptive experiences linked to anticipation of imminent use can function to reinforce smoking. They encourage future research to better “dissect the cognitive-attentional and affective processes that lead up to a relapse (i.e., anticipation)…” (p. 133). Consistent with this aim, in the present study, we tested the impact of anticipation of imminent use on the degree to which smokers choose to engage smoking cues.
Most lab-based drug cue reactivity studies involve participants who do not expect to use the drug during the experiment (see Bradford et al., 2015). Outside the lab, however, responses to drug cues occur in conjunction with a variety of expectations regarding drug use. Understanding of the etiological importance of cue reactivity must account for these different expectations. This study used novel stimuli and recruited both continuing and treatment-motivated smokers to evaluate the impact of perceived smoking opportunity on smoking anticipation experiences, and in particular on the manner in which smokers experience both smoking and nonsmoking stimuli.
Although we believe research manipulating treatment seeking status and perceived smoking opportunity is vital, it is unsurprising that systematic investigations rarely are conducted. It can be difficult to recruit smokers who wish to quit when advertising for smokers more generally. In some instances smokers initially claim an interest in quitting that does not translate into specific actions to do so. This change of heart may reflect genuine ambivalence about quitting or perhaps a desire to gain entrance into a paid experiment by reporting what participants hope will be the right response. Alternatively, because motivation to quit is not a static phenomenon, interest in quitting may shift from the time of initial assessment to actual participation in the experiment. Regardless of the explanation, in the present study we sought to remedy this concern by recruiting motivated to quit participants by having them initially visit a quit-smoking clinic.
We also believe that it is challenging to create optimal circumstances to experience anticipation of imminent use in the lab. On one hand, there must be some base level of desire to smoke. If participants are told to smoke at the start of, or just before entering, the experiment, for example, there may be little desire or anticipation to smoke another cigarette during the experiment (see Wilson & Sayette, 2015). On the other hand, if there is too much withdrawal due to an extended period of abstinence, then even modest delays prior to consumption can be experienced unpleasantly (Sayette et al., 2003). While we believe that in the natural environment these two conditions are often met, it is trickier to execute them in the lab using standardized approaches to smoking abstinence. Following interviews with 21 smokers of varying ages and levels of dependence, we opted to use a 5-hr abstinence period, a duration less than what is used in nearly all studies requiring nicotine deprivation, but still we hoped long enough to lead smokers to want to smoke.
Finally, calibrating the delay phase in the laboratory is difficult. It must be long enough to permit savoring but not too long that frustration ensues. For instance, in past work we found that delaying smoking for even short durations turned affect-related facial responses from predominantly positive to predominantly negative (Sayette et al., 2003).
In sum, continuing smokers and those seeking treatment at a local smoking clinic participated in a smoking cue exposure study following 5-hr of smoking abstinence. Smokers were randomly assigned to smoking opportunity condition (i.e., told they either would or would not be permitted to smoke during the experiment). They were assessed using traditional self-report affect and urge measures during an in vivo smoking cue exposure task and also completed two tasks designed to evaluate the attractiveness of smoking and nonsmoking stimuli. (Because it is impractical to randomly assign smokers to treatment-seeking status, efforts were undertaken to try to keep the groups to approximately the same age and gender composition, and when necessary to adjust statistically for such group differences in status.) We hypothesized that participants who expected to smoke would generally find smoking and nonsmoking stimuli to be more pleasant than would those not expecting to smoke and that this observation would be particularly true for continuing smokers who expected to smoke. Thus, we predicted a treatment motivation by smoking instruction interaction. Regardless of the presence of an interaction, however, we predicted, in accord with the findings from Wilson et al. (2012), that continuing smokers expecting to smoke soon would be more likely than each of the other three groups to find the experimental stimuli rewarding.
Our main goal was to examine the impact of (a) treatment-seeking status (status) and (b) the perceived opportunity to smoke soon (opportunity) on smokers across four response domains. We recruited participants for the motivated to seek treatment (MST) condition who had first visited a quit smoking clinic offered by a University of Pittsburgh-affiliated hospital. Smokers (responding to newspaper advertisements) were considered to be currently not interested (CNI) in quitting if they reported not being interested in seeking treatment or cutting down on their cigarette use within the next 30 days (see also Geier, Sweitzer, Denlinger, Sparacino, & Donny, 2014). Importantly, as noted above, while MST smokers did intend to quit smoking soon, they had not yet begun to reduce cigarette use. Thus, they were seriously interested in quitting but in most cases not yet ready to make an active attempt. This study used a factorial design, with MST and CNI smokers, all of whom were required to abstain from smoking for 5-hr prior to entering the laboratory, being randomly assigned to receive instructions indicating that they either would (Yes) or would not (No) have the opportunity to smoke during the experiment. Thus, there were four groups of participants: MST-Yes (N= 53); MST-No (N= 54); CNI-Yes (N= 60); and CNI-No (N= 60).
Participants consisted of 227 smokers (117 female, 110 male) aged 30–70. (To reduce potential age differences between the two types of smokers, we excluded participants under 30 years old, who tended to be uninterested in quitting.) All participants were required (and reported) smoking ≥ 10 cigarettes per day for at least 12 continuous months. To assess exclusionary criteria, participants completed an initial telephone interview and attended a screening session. Participants were excluded if they reported a medical condition that contraindicated nicotine or if they were illiterate. Informed consent was obtained from all participants. Based on self-reported assessment, the final sample was 45% African-American, 46.5% Caucasian, 7% multiracial, 1%, Asian, and 0.5% American Indian.
Smokers who responded to ads or who had been informed about our study after visiting a smoking clinic underwent a phone interview to exclude those not meeting selection criteria. Smokers who met selection criteria were asked to attend a 30-min screening session to determine whether they would qualify for a 2-hr experimental session. Smokers were told that they would be paid $5.00 at the screening session if they were determined to be ineligible or $60.00 at the experimental session if they were determined to be eligible.
The 30-min screening session was held at either our laboratory or the smoking clinic. At this session, participants were required to show photo identification, to provide a baseline CO reading, and to complete a brief interview regarding their current medical status and medication usage. Participants were ruled eligible if they were aged 30–70, their CO reading was >10 ppm, and their current medical status and medication usage did not indicate that smoking a cigarette would be especially harmful to them. Ineligible participants were paid $5.00 and permitted to leave. Eligible participants were provided with instructions for the experimental session. Specifically, they were told to bring a pack of their preferred brand of cigarettes and to abstain from smoking for 5 hours prior to the experimental session. They also were told that breath samples would be collected to ensure that they had abstained from smoking for 5-hr.
Table 1 presents a timeline of the procedures. Sessions were begun between 10:00 a.m. and 2:00 p.m. To check compliance with smoking abstinence instructions, participants reported the last time they smoked and provided a CO reading. Participants had to have a CO that did not exceed 20 ppm. [This value is higher than in some of our past studies (e.g., Sayette & Hufford, 1994; Sayette et al., 2001; Sayette, Loewenstein, Griffin, & Black, 2008), which reflects an abbreviated abstinence duration requirement.] Participants presented their pack of cigarettes and lighter to the experimenter and completed an assessment of current mood and desire to smoke. Using a scale with 1 = strongly disagree and 7 = strongly agree, current mood was measured using two items (I am happy, joyful, or pleased; I am depressed, angry, worried, or frustrated) (Carter & Tiffany, 2001), and desire to smoke was measured by one item (I would like to smoke a cigarette). Composite baseline mood reflected the difference between the negative and positive item. These three items were embedded within a 6-item current experiences scale that also included three filler items designed to make the focus on mood and smoking desire less salient.
Depending on opportunity condition, participants next randomly received one of two sets of instructions regarding the opportunity to smoke during the study. Participants in the Yes condition were told:
“People often want to know whether they will be able to smoke during this experiment, and YES, during that break you would be welcome to smoke a cigarette. If you choose to have a cigarette during this smoke break, you’ll be able to do so in this room, as it has been approved for that purpose and has an air-clearing machine and special ventilation in the ceiling. Please note that this break is scheduled into the study, regardless of whether you choose to smoke, and does not affect how long you will be here.”
Participants in the No condition were told:
“People often want to know whether they’ll be able to smoke during this experiment and unfortunately the answer is no. Do you think you’ll be okay with not smoking during the 2 ½ hours in which you’re here? [All indicated they would be okay not smoking while in the experiment.] Participants received one of two consent forms corresponding to their condition, stipulating that they either would (Yes) or would not (No) be allowed to smoke during the experimental session. We previously used this approach to reinforce instruction manipulations (e.g., Sayette et al., 2003). Equivalent numbers of MST and CNI smokers were assigned to each condition.
After the opportunity manipulation, participants (1) completed measures of urge and affect; (2) viewed the smoking/alcohol images; (3) listened to a 5-min clip of a comedy routine; and (4) underwent an in vivo smoking cue exposure during which urge and affect were again recorded.
Following opportunity instruction, participants reported urge to smoke (post-opportunity urge) using a rating scale ranging from 0 (labeled “no urge to smoke at all”) to 100 (labeled “the most intense urge to smoke that you have ever felt.” (Juliano & Brandon, 1998; Sayette et al., 2001). Participants then indicated their current affect state using the Positive and Negative Affect Schedule (PANAS: Watson, Clark, & Tellegen, 1988).
Participants rated a series of images presented on a computer monitor that included nine alcohol and nine cigarette ads (interspersed) purportedly to be used in a future study. (Images were located on the internet and selected for inclusion based on ratings obtained in a pilot study in our laboratory.) We chose to use alcohol images as our comparison stimuli as we felt it would provide an especially stringent test of our hypothesis that anticipating smoking a cigarette would make proximal smoking cues an especially alluring stimulus, even when compared to other consummatory substances that are often used by smokers. Second, the name of the laboratory includes both alcohol and smoking and consequently we believed it added plausibility to our cover story that we would want participants to evaluate these images for future research projects. Participants viewed each image for as long as they wished before hitting the space bar to advance to the next image, and rated each image using a 1–9 scale, where 1 = not at all pleasant and 9 = extremely pleasant. To test the reactions to cigarette images (relative to alcohol images), we assessed: 1) the rating assigned to each image (Pleasantness) and 2) the amount of time participants chose to view each image (Time2).
Participants next listened to a 5-min comedy clip, which was designed to serve as our non-smoking rewarding stimulus, taken from a stand-up act performed by comedian Jerry Seinfeld. None of the jokes included sexual or aggressive content. Following the clip, participants rated the routine using two 0–10 scales with 0 = to not at all and 10 = extremely to indicate how entertaining and how funny they found the comedy clip to be.
Following the comedy routine and a 2-min rest period, participants completed an in vivo smoking cue exposure manipulation that has been used previously in our lab (e.g., Sayette & Hufford, 1994; Sayette et al., 2001; Sayette & Parrott, 1999). A covered tray was placed on a desk in the experimental room in front of the participant. Participants were told not to touch the tray. Once the experimenter had returned to an adjacent room, participants were instructed over an intercom to remove the cover from the tray, revealing the cigarette pack they had brought to the session along with a lighter and ashtray. (Participants’ own cigarettes served as the smoking cue to increase magnitude of reactivity.) They were instructed to remove a cigarette and light it without putting it in their mouths. Participants next were told to put down the lighter, hold the cigarette comfortably, and look at it without placing it in their mouths. After 15-sec participants again rated their urge to smoke (in vivo urge) and completed a PANAS rating (in vivo PANAS). Seven seconds later, participants extinguished their cigarette in the ashtray.
Following in vivo cue exposure, participants completed three questionnaire packets. The first included a form assessing demographic variables as well as several other questionnaires3. The second packet included a modified version of the smoking history assessment developed by Shiffman and colleagues (Shiffman, Paty, Kassel, Gnys, & Zettler-Segal, 1994), as well as the Nicotine Dependence Syndrome Scale (Shiffman, Waters, & Hickox, 2004), revised Fagerström Test for Nicotine Dependence (Heatherton, Kozlowski, Frecker, & Fagerstrom, 1991), and Smoking-Specific Felt Attitudinal Ambivalence Scale (Lipkus, Green, Feaganes, & Sedikides, 2001). The third packet included the Contemplation Ladder (Biener & Abrams, 1991) and a form asking participants about the study’s purpose. Once participants completed all three packets, they were debriefed, paid $60.00, and permitted to leave.
Table 2 presents participant characteristics by group. Groups were equivalent on ethnic and racial composition, years of schooling (M= 13.2), gender distribution (51.5% female), number of cigarettes/day, and baseline mood. Groups did not differ on CO readings at study outset.
Participants randomly assigned to the Yes and No smoking opportunity conditions did not vary on age, FTND, motivation to quit (using the contemplation ladder) or wanting to smoke a cigarette. As expected, MST participants reported significantly higher scores on the contemplation ladder (M= 8.52) than did CNI’s (M=5.63), F (1, 206) = 67.76, p <.0001. At the outset of the experimental session, MST’s reported lower scores for wanting to smoke a cigarette (M=5.35) than did CNI participants (M=6.28), F (1, 222) = 17.19, p <.0001. MST smokers also were slightly older (M= 50.5 years) than were CNI smokers (M= 48.3 years), F (1, 221) = 4.02, p <.05. MST participants reported lower FTND values (M= 4.6) than did CNI participants (M= 5.3), F (1, 219) = 5.94, p <.02. Except when otherwise noted, our general approach was to covary for age, FTND, and baseline mood (which could affect the DVs of interest).
A 2 (status) × 2 (opportunity) ANCOVA with age, FTND, baseline mood, and baseline desire to smoke as covariates, revealed a main effect of status, F (1, 211) = 8.67, p <.004 (d=.40). Specifically, urge to smoke scores were significantly higher for CNI’s (M= 65.38, SD = 23.50) than for MST’s (M=55.97, SD = 23.64). No other effects were significant.
Scores on the positive and negative subscales of the PANAS – covarying for age, FTND, and baseline mood – did not vary significantly across conditions (p’s >.15).
We examined two variables related to the experience of viewing smoking cues: pleasantness of the smoking images relative to the alcohol images and time spent viewing the smoking relative to the alcohol images. [In each case the smoking image values were responses to smoking images subtracting values for the alcohol images.]
A 2 (status) × 2 (opportunity) ANCOVA with age, FTND, and baseline mood as covariates, revealed a main effect of status, F (1, 214) = 4.33, p <.04 (d=.28), and a marginal effect for opportunity, F (1, 214) = 2.97, p <.09. Pairwise least square means tests revealed that as expected, the CNI-Yes smokers enjoyed viewing cigarette images significantly more than did each of the other three conditions (all p’s <. 04), all d’s ranging from .39 to .51).4 (See Figure 1.)
We next examined the amount of time that participants chose to view each slide. No main effects emerged, but there was a marginally significant interaction, such that the two groups that might have a conflict between their quitting motivation status and their perceived opportunity to smoke (MST-Yes and the CNI-No smokers) tended to spend relatively less time viewing the smoking slides compared to the nonconflict groups (MST-No and the CNI-Yes smokers) F (1, 214) = 3.10, p <.08.
Taken together, the CNI-Yes smokers found viewing smoking images to be significantly more pleasant than did smokers in the other three conditions and (along with the other nonconflict group) tended to look at the smoking images for relatively longer periods of time.
In addition to testing the impact of condition on responses to smoking cues, we investigated whether anticipating smoking would enhance the experience of generally pleasant stimuli unrelated to smoking. Unfortunately, only 57% of participants reported that they would recommend the comedy clip to a friend, a value which did not differ by condition. It may be that the clip was not sufficiently enjoyable to many of our participants.
The two comedy measures (entertaining, funny) were highly correlated, r (227) =.91, p <.0001. We therefore created a composite ‘enjoy’ variable and conducted a 2 (status) × 2 (opportunity) ANCOVA with age, FTND, and baseline mood as covariates, and enjoy as the dependent variable. There was a main effect for age, such that older participants liked the clip more than did younger participants, F (1, 214) = 4.95, p <.03. Although in the direction of an effect for opportunity, such that the CNI-Yes participants liked the clip more than did those in the other three groups], no effects reached significance (p’s > .12).
During exposure to the lit cigarette, we computed a 2 (status) × 2 (opportunity) ANCOVA with age, FTND, baseline mood, and baseline desire to smoke as covariates, and in vivo urge to smoke as the dependent variable. As expected, baseline desire to smoke was a significant covariate, F (1, 213) = 34.34, p <.0001. There also was a main effect of status, F (1, 213) = 3.95, p <.05 (d=.27), such that even after adjusting for baseline levels of desire, CNI participants reported greater urge to smoke (M= 73.82, SD = 28.82) than did MST (M= 66.01. SD = 28.84) participants.
We computed a 2 (status) × 2 (opportunity) ANCOVA with age, FTND, and baseline mood as covariates, and the in vivo PANAS as the dependent variable. (This measure was introduced ¼ of the way through data collection and thus was completed by 165 participants.) For the negative subscale, as expected, baseline mood was a significant covariate, F (1, 153) = 4.76, p <.04. There also was a main effect of opportunity, F (1, 153) = 5.82, p <.02 (d=.33), such that even after adjusting for baseline mood, Yes participants reported a less negative mood (M = 13.93, SD = 7.41) than did No (M = 16.40, SD = 7.83) participants. For the positive subscale, though the CNI-Yes group reported the highest mean value among the four groups, none of the differences reached significance (p’s > .07).
Craving is often discussed in the clinical literature and among laypersons as if it is a unitary phenomenon. Yet cravings may manifest quite differently depending on the context in which they arise (Sayette, in press). While most discussion of cigarette craving focuses on frustrating experiences of those trying to quit that precede relapse (Tiffany 1992), there also are important conceptual and clinical reasons to examine cravings associated with the anticipation of imminent smoking in continuing smokers. The present study employed a diverse set of measures to examine two contextual variables (current interest in quitting and perceived opportunity to smoke soon) thought to affect the degree to which imminent smoking is anticipated.
We were particularly interested in determining whether continuing smokers anticipating the opportunity to smoke very soon would be most likely to find smoking stimuli (separate from the act of smoking itself) to be attractive and enjoyable to observe. Results provide some initial support for this hypothesis. Smokers who were not trying to quit and who were informed they would be able to smoke imminently in our study seemed to find the smoking images relative to alcohol images5 to be more pleasant than did the other groups of smokers. They also tended to report more positive affect following a traditional in vivo smoking cue exposure manipulation. In tandem with recent findings indicating that anticipation of imminent use dampened response to a stressor (Bradford et al., 2015), the present data suggest that certain preconsumptive factors (such as anticipation) can, independent of actual consumption, improve emotional experiences and enhance the attractiveness of the smoking cues that trigger cravings.
These data offer support for incentive motivation theory, which emphasizes the drive to consume a drug. The present data also suggest, however, that under certain conditions the cues themselves become attractive. One might imagine that CNI-Yes smokers would be eager to race through the smoking image task knowing that the sooner they complete the tasks the sooner they can smoke. From this perspective, the smoking images are merely an obstacle interfering with smoking. Yet the data suggest that these smokers were not in a hurry to do so, and instead seemed to find the smoking images and the in vivo cue exposure pleasant to view (see also Wilson et al., 2012). In some cases drug users may wish to delay actual use in order to prolong pleasure associated with anticipation of imminent use [i.e., relish the moment (Kavanagh, Andrade, & May, 2005)]. These findings reinforce prior animal research suggesting that in certain situations, drug cues themselves may become desirable (e.g., Flagel et al., 2011; Palmatier et al., 2007). Moreover, such cues may be more likely to assume rewarding properties in certain animals (sign trackers) than in others (goal-trackers) (Yager & Robinson, 2015), suggesting a potential individual difference variable to be evaluated in future studies with humans.
The present findings also suggest that a smoker’s motivation to seek treatment is a factor that influences craving intensity. Both following the smoking opportunity instruction and during a subsequent in vivo smoking cue exposure manipulation there was a main effect of motivation to seek treatment status, such that urge to smoke was greater for CNI than for MST smokers. This finding, which emerges from a sample that contrasted both types of smokers within one study is in accord with the conclusions from a review of prior studies that tended to use either continuing or treatment-seeking participants, but not both (Wertz & Sayette, 2001a).
We had expected that status and opportunity also would combine to affect the degree to which a comedy clip was experienced as enjoyable. While there were some effects in the expected direction, none reached significance. Thus, we failed to support the hypothesis that being in a pleasant anticipatory state would lead to enhanced reward associated with other nonsmoking stimuli. Because nearly half the participants did not seem especially amused by the comedy clip (reporting that they would not recommend it to a friend) we wonder if a more engaging clip might have offered a better test.
We believe that the present study, which included one of the larger samples of nicotine-deprived smokers found in the smoking cue reactivity literature, offers a reasonable, if not overwhelming, level of evidence to suggest that smokers not interested in quitting and who anticipate smoking very soon will find smoking cues attractive. We hope that our findings are viewed as providing preliminary support for the idea that responses to smoking cues sometimes can have pleasant features, and that our CNI-Yes group tended to manifest distinct response patterns from the other groups of smokers. More importantly, we hope that these findings will stimulate additional research, especially when one considers that they emerged despite the challenges associated with capturing this pleasant anticipatory experience in the laboratory.
First, as noted above, it is difficult to specify an optimal duration of smoking abstinence in order to test smokers in the lab who are nicotine-deprived enough that they wish to smoke, but are not so deprived that any extra delay (to complete the measures) is going to be experienced negatively. We aimed to model in our study the type of anticipatory experience a diner might enjoy while awaiting the arrival of their favorite dish at a wonderful restaurant, including stealing glances at a diner at a nearby table who is already enjoying the same selection. That pleasant experience disappears if s/he has just eaten a large meal before entering the restaurant, or conversely, if s/he has gone too long without eating, such that any further delay before the prized dish is served becomes unpleasant. We conducted pilot testing prior to selecting a 5-hr abstinence duration and found it to be preferred to an individually tailored abstinence approach (it is hard to determine in advance the ideal abstinence interval length for each person). Indeed, mean urge to smoke levels – which were above the scale midpoint but lower than in many prior smoking studies with longer abstinence requirements (see Wertz & Sayette, 2001a) suggest that we were at least close to the right abstinence interval. Nevertheless, surely there were smokers in our study who would have been more likely to have experienced a pleasant anticipatory craving experience if the delay interval had been either longer or shorter. Moreover, it appears that a 24-hr smoking abstinence period did not prevent the emergence of a viable smoking anticipation state (Bradford et al., 2015), so clearly more investigation of this issue is indicated.
Second, it can be challenging to recruit smokers who wish to quit smoking very soon. Motivation to quit is better viewed as a dimensional than a dichotomous construct that can shift over time. Based on our prior struggles to recruit MST smokers, in this study we recruited these participants after they had attended a smoking cessation clinic to discuss quitting. But even in this setting the vast majority of the smokers ultimately never reached quit day before breaking off contact with the clinic. Thus, the MST smokers in the present study may have been less motivated to quit than we would have hoped. Accordingly, the observed differences between MST and CNI participants, though clearly significant based on contemplation ladder scores, nevertheless might have led to an underrepresentation of the true magnitude of the effect of treatment motivation on craving.
Finally, it is tricky to manipulate perceived smoking opportunity in the lab. Ideally an experiment could capture the pleasant anticipation experience described in the restaurant scene above and then administer a battery of tests during that experience without shattering this pleasant, though fragile, moment altogether. We observed just how difficult it is to assess pleasant anticipatory cravings in a prior study (using a longer abstinence requirement than in the present study) in which even slightly varying the delay interval before smoking (from 15-sec to 60-sec) altered the affective experience (Sayette et al., 2003, experiment 2).
In sum, we believe pleasant anticipatory cravings are common to continuing smokers in the real-world while they savor the moments prior to a regularly scheduled smoke break. Nevertheless, the constraints of a laboratory setting create challenges to unobtrusively capturing them. We believe that this study, as well as our prior neuroimaging study (Wilson et al., 2012) and the study by Bradford et al. (2015), suggests that such imminent anticipatory states can be elicited and observed in a controlled setting, however, and moreover, that the study of this form of drug anticipation merits further inquiry. These few studies by themselves do not establish directly that anticipation of imminent smoking is rewarding, however, and future research is needed to evaluate the impact of this state on positive and negative moods across a variety of contexts. Our findings also raise questions regarding a potential conceptual distinction between the pleasantness of smoking cues and the pleasantness of anticipatory craving, per se. It would be valuable to show that anticipation of imminent use is actually reinforcing, in that smokers would be willing to work (or pay) to receive such moments of anticipation. In addition to experimental investigations, complementary research using ecological momentary assessment (Shiffman & Kirchner 2009) would seem to be especially valuable for capturing the individualized routines leading to pleasant anticipatory cravings.
There is a chasm between how cravings are regularly experienced by smokers in the natural environment and how they are studied in the laboratory. If cravings in the continuing user offer regularly scheduled anticipatory experiences that also may be accompanied by pleasant reactions to drug cues (as found in the present study), then it follows that quitting a drug necessarily leads to the abrupt loss of these moments, and their loss may help explain the increase in negative affect following cessation (Baker et al., 2004b; 2006).
If replicated, this research (along with Bradford et al., 2015 and Wilson et al., 2012), may offer a new research direction that considers the anhedonic effects of quitting due to the loss of these rewarding anticipatory cued experiences (Leventhal, Waters, Kahler, Ray, Sussman, 2009). Moreover, the present data suggest the potential efficacy of cognitive-behavioral or mindfulness training interventions aimed at managing the loss of these anticipatory states (independent of actual drug ingestion experiences) after quitting. Treatment often focuses on the times of day that shift from drug use to nondrug use upon quitting. Less attention is paid to the period just prior to drug use, which ostensibly (and pharmacologically) remain unchanged. Indeed, what may have been a pleasant or less stressful period linked to anticipating the next smoke break or time at the bar now may turn somber or more anxious, with nothing to look forward to or to eagerly anticipate. This affective downshift may be all the more threatening due to smokers’ lack of insight regarding their cravings (Sayette et al., 2008).
Conceptually, focus on anticipation of imminent use also raises questions about how these anticipatory states develop over the course of addiction. Robinson and Berridge (1993) posit that drug wanting sensitizes with repeated use, while drug liking does not. It is unclear if pleasant anticipation states become sensitized, like wanting, or instead become muted over time, like the hedonic experience of the drug. Research that also recruits light smokers who have just begun smoking would provide optimal conditions to evaluate potential differences in sensitivity to anticipation of imminent use. Moreover, research would be useful to evaluate interventions that focus on the loss of these anticipatory experiences after quitting and to develop strategies not just to add replacement rewards into their lives, but to ensure these new rewards create pleasant anticipation states (e.g., have PDAs signal 30-min prior to meeting friends). Clearly continued research on the impact of contextual factors such as motivation to quit and perceived smoking opportunity will lead to a more nuanced understanding of drug cue reactivity and craving.
This research was supported by NIH (National Cancer Institute) grant R01 CA159463-12 to Michael Sayette. We thank George Loewenstein for his helpful comments throughout the study, Betty Marconi for her help with recruitment, and Michael Laurent, Mike Sayers and the students and staff at the Alcohol and Smoking Research Laboratory for assistance collecting data.
1Because smokers sometimes can have cigarettes available but still not perceive an opportunity to smoke – e.g., orthodox Jews on the Sabbath may have cigarettes available yet “know” that they will not smoke (Dar, Stronguin, Marouani, Krupsky, & Frenk, 2005), we instead use the term perceived smoking opportunity herein to refer to the perception that the cigarette is able to be smoked (see Wertz & Sayette, 2001a).
2This measure was adapted from work linking image viewing time and attraction (Field, Mogg, & Bradley, 2004; Lykins, Meana, & Strauss, 2008). Because we worried that too much delay due to repeated assessment of emotion items might undermine our fragile anticipation of imminent use phenomenon (Sayette et al., 2003), we assessed only image pleasantness, assuming a “not at all pleasant” rating would approximate an unpleasant reaction.
3We included the following questionnaires: the Snaith-Hamilton Pleasure Scale (Snaith, Hamilton, Morley, Humayan, Hargreaves, & Trigwell, 1995), Affective Intensity Measure (Larsen & Diener, 1987), Balanced Inventory of Desirable Responding-6 (Paulhus, 1991), NEO Five-Factor Inventory (Costa & McRae, 2003), Barratt Impulsiveness Scale (Patton, Stanford, Barratt, 1995), and Mindful Attention Awareness Scale (Brown & Ryan, 2003). Because they were unrelated to our outcomes and to conserve space we do not address them further in this report.
4Although common in the literature, using covariance analyses in a quasi-experimental design is suboptimal. We are unaware, however, of a straight-forward solution beyond our selection of similar gender compositions and an age range that we hoped would constrain potential preexisting differences between the CNI and MST groups (e.g., to prevent having a disproportionate number of young continuing smokers). Note also that all of these significant contrasts remain when covariates are removed from the analysis.
5In hindsight it would have been useful to have assessed participants for alcohol misuse.