Cue-induced cigarette craving and mixed emotions: A role for positive affect in the craving process

Abstract

Craving is an important component of nicotine addiction, and extant research has demonstrated a clear link between cue-induced craving and negative affect, with mixed results in the positive affect domain. The current study was designed to test the idea that cue-reactive craving might be associated with a mixed emotional process, or the simultaneous experience of positive and negative affect. Participants were 86 non-deprived regular smokers and tobacco chippers who provided simultaneous ratings of positive and negative affect during cue exposure to pleasant, unpleasant, neutral and cigarette cues. Results indicated that self-reported craving was elevated in response to cigarette cues compared to other valenced cue types and craving was higher to pleasant cues than either neutral or unpleasant cues. Mixed emotional responses were higher to cigarette cues than other cue types. In addition, mixed emotional responses to cigarette cues predicted craving even after controlling for smoker type, difficulties regulating negative emotion, baseline craving level and mixed emotional responses to neutral cues. As the first study to investigate mixed emotions and cigarette craving, our results highlight the importance of examining the relationship between cue-reactive craving and emotional response using models of emotion that allow for measurement of nuanced emotional experience. In addition, our findings suggest that positive affect processes may indeed play a role in craving among non-deprived smokers.

Introduction

Craving for cigarettes, whether induced by withdrawal or in response to cues, has been clearly linked to affective processes (Tiffany, 2009). In the realm of cue-reactivity, a robust relationship has been demonstrated between negative affect and craving (Brandon et al., 1996, Carter and Tiffany, 2001, Drobes and Tiffany, 1997, Maude-Griffin and Tiffany, 1996, Taylor et al., 2000). Despite these findings, the precise nature of the relationship between craving and affect remains uncertain. Further examination of the affective processes associated with cue response may prove important, particularly if affective responses contribute to self-reported craving and/or behavioral outcomes.

The relationship between cue-reactive craving and negative affect is clearly established in the literature. A host of evidence suggests that negative affect manipulations, even without the addition of smoking cues, prompt increased craving levels (Maude-Griffin and Tiffany, 1996, Taylor et al., 2000). Negative affect predicts craving when measured prospectively (Brandon et al., 1996, Drobes and Tiffany, 1997), and negative states show robust correlations with craving (Carter and Tiffany, 2001, Lee et al., 2007, Singleton et al., 2003). Moreover, presentation of smoking cues tends to increase negative affect more than presentation of neutral cues (Conklin et al., 2010, Doran et al., 2008).

At first blush, it appears that positive affect typically either has no correlation to smoking urge (e.g. Drobes and Tiffany, 1997, Kassel et al., 2007), or is associated with decreased urge (Bailey et al., 2010, Maude-Griffin and Tiffany, 1996, Taylor et al., 2000, Tiffany, 2009).

Theoretically, however, positive affect has been linked to urge. A recent model of impulsivity suggested that certain individuals are likely to act rashly (e.g. succumb to urges) when in a positive mood (Cyders and Smith, 2008, Cyders et al., 2007). In addition, Baker and colleagues (Baker, Morse, & Sherman, 1987) proposed a craving model with two separate urge pathways such that a positive-affect urge network may be activated by appetitive stimuli, particularly for non-deprived smokers. As an example, in a study comparing deprived and continuing smokers on craving and affect ratings over a 24-hour period, Zinser and colleagues (Zinser, Baker, Sherman, & Cannon, 1992) found a significant positive relationship between positive affect and craving for continuing smokers, whereas deprived smokers expressed concordant negative affect and craving (however, see Brandon et al., 1996 for contradictory findings).

Several of the studies that found a negative relationship between positive affect and craving have investigated the effect of positive mood, typically induced in a laboratory setting, on subsequent cue-reactive craving (Maude-Griffin and Tiffany, 1996, Taylor et al., 2000). Whether mood state influences reaction to cues is a different, albeit related, question than whether self-reported craving increases in response to valenced cues or if changes in affect follow from presentation of cigarette cues. Self-reported craving in response to valenced cues has received limited empirical attention in the realm of cue-reactivity, as the typical cue-reactivity study compares craving responses to cigarette cues with craving responses to neutral cues (Carter & Tiffany, 1999). However, recent work comparing craving responses to a full set of valenced cues (e.g., including both pleasant and unpleasant cue stimuli) found heightened craving in response to positive cues when compared to negative and neutral cues (Muñoz et al., 2010, Robinson et al., 2011). Moreover, in the realm of psychophysiology, startle eyeblink responses (where heighted responses are indicative of negative affect) are dampened to both cigarette and pleasant cues (Cui et al., 2012, Dempsey et al., 2007). Overall, these studies suggest that smoking cues are appetitive.

Other recent work has examined changes in affect following the presentation of cigarette cues, with mixed results. One study found that generalized positive affect decreased and negative affect increased in response to cigarette cue exposure (Doran et al., 2008). Another study found that smoking cues are associated with higher excitement (a specific positive feeling), when compared to non-smoking cues (Conklin et al., 2010). A third recent study examining self-reported mood in response to valenced and cigarette cues found that positive mood was higher to cigarette and positive cues compared to neutral and unpleasant cues (Robinson et al., 2011). In sum, although there is burgeoning evidence that smoking cues are appetitive, and theoretical frameworks proposing a link between positive affect and craving, the remaining empirical research on the relation between positive affect and craving is mixed. Considering that a synonym often used for craving is “desire,” which has a positive connotation, the mixed evidence is an apparent conundrum. Research focused on positive affect and craving may benefit from considering alternative approaches and methodologies to determine if, when, and for whom positive affect relates to craving.

An additional piece of the puzzle is that positive and negative affect are rarely compared to one another. For example, in a study comparing affective responses to cues, positive affect was lower in response to an imagery exposure with urge content compared to an imagery script with neutral content (Drobes & Tiffany, 1997). However, ratings of positive affect across the entire study were considerably higher than ratings of negative affect, a point not raised by the authors, as they focused on comparisons of smoking to neutral cues rather than comparisons of negative and positive affect.

Examining changes in emotional response to cues is important, but comparison of levels of negative and positive affect may provide different—and potentially equally important—information. For example, Sayette and Hufford (1995) found that participants manifested more positively valenced facial expressions than negative during initial cue exposure, and that more positive urge characteristics were self-generated during cigarette cue exposure compared to neutral cues, a relationship not found for negative urge characteristics (Sayette & Hufford, 1997). This study suggests that comparison of negative and positive affect might highlight the role of positive affect in cue-reactivity which is often obscured in other study designs. Moreover, a comparison of emotional responses may elucidate the relative contributions of positive and negative affect, both together and separately, on craving responses.

Highlighting the role of positive affect in cue-reactivity is only important if positive affect influences craving and/or behavior. Considering the robust relationship between craving and negative affect, it may be that “pure” positive affect has little role in the craving process. However, there may be benefit in modeling simultaneous positive and negative affect, dubbed “mixed emotions”(Larsen and McGraw, 2011, Larsen et al., 2001, Larsen et al., 2004, Larsen and Stastny, 2011). In basic emotion research, Larsen and colleagues have demonstrated that many people experience heightened positive and negative affect when confronted with a film that is both funny and tragic; when experiencing the celebration and fear associated with graduating from college; or when reacting to disappointing wins and relieving losses (Larsen et al., 2009, Larsen et al., 2001, Larsen et al., 2004).

The concept of mixed emotions calls into question the assumption that positive and negative affect must be mutually exclusive, an assumption that is common in basic emotion research (e.g., Russell, 2003) as well as implicitly endorsed in many of the cue-reactivity studies reviewed above. If negative affect and positive affect are opposites, and mutually exclusive, when negative affect goes up, positive affect must, by definition, go down. However, if positive and negative affect are independent and modeled simultaneously, we can establish if cues elicit relatively pure emotional responses or if cue-reactivity may be associated with a mixed emotional response.

The idea that craving might involve both positive and negative affect processes is certainly not new (Baker et al., 1987, Breiner et al., 1999, Kavanagh et al., 2005). However, to our knowledge, the association between craving and mixed emotion has not been explicitly tested. Nonetheless, several studies have yielded findings consistent with a mixed emotional perspective. Taylor et al. (2000) created a variety of imagery scripts that differed in valence and smoking urge, and examined how the scripts influenced positive affect, negative affect and craving. For positively valenced scripts, the inclusion of urge-related material decreased positive affect and increased negative affect compared to the no-urge scripts. However, for negatively valenced scripts, the inclusion of urge-related material resulted in lower negative affect and heightened positive affect compared to no-urge negative scripts. Thus, the addition of the urge component attenuated the effect of negative stimuli on self-reported affect. These results are consistent with a mixed emotional experience; the urge component elicited some positivity to the negative scripts, and some negativity to the positive scripts.

One recent study investigating ambivalence about smoking explicitly examined mixed emotional facial expressions during a cue reactivity paradigm. Among regular smokers who were 7 hours deprived, Griffin and Sayette (2008) measured positive and negative facial expressions while smokers were exposed to smoking and neutral cues. They found that 24% of participants exhibited both positive and negative facial expressions in response to smoking cues, where no mixed expressions (termed ‘ambivalent’ by the authors) were evident in response to neutral cues. Moreover, participants who displayed mixed expressions also evidenced elevated withdrawal symptoms and a greater interest in quitting compared to participants who did not display a mixed emotional response.

It is important to test the relationship between craving and mixed emotions for several reasons. A mixed emotional experience is qualitatively different, and appears to be theoretically less common than either a “pure” negative experience or a “pure” positive experience (Larsen, Hemenover, Norris, & Cacioppo, 2003). In other words, there is inherent conflict in a mixed emotional experience that may result in experiential discomfort. If craving elicits a mixed emotional experience, and subsequent cigarette smoking resolves the mixed feeling, simultaneous positive and negative reinforcement processes may be at work. Moreover, considering mixed emotional responses allows a role for positive affect in the craving process, which is theoretically suggested and may be obscured by considering positive and negative separately.

The current study was designed to extend cue-reactivity research in several ways. First, in addition to providing smoking and neutral cues, as is common in cue-reactivity studies (e.g., Bailey et al., 2010, Doran et al., 2008, Sayette and Hufford, 1995), pleasant and unpleasant affective cues were also included in order to better facilitate understanding of the craving-emotion relationship. Specifically, pleasant and unpleasant cues were included as additional comparison points to neutral cues to replicate recent findings that positive cues prompt heightened craving compared to negative and neutral cues (Muñoz et al., 2010, Robinson et al., 2011).

Second, we investigated the relationship between mixed emotion and craving using a measure of emotion developed from an emotion model that posits negative and positive affect as partially independent systems (Evaluative Space Model; Cacioppo and Bernston, 1994, Cacioppo and Berntson, 1999). We predicted that compared to the other valenced cue types, mixed emotions would be heightened in response to cigarette cues. We also predicted that mixed emotions would predict self-reported craving above and beyond mixed emotional responses to neutral cues.

Third, to attend to critiques that cue-reactivity research often neglects behavioral outcomes (Perkins, 2009, Rosenberg, 2009, Sayette et al., 2000), we hypothesized that craving in response to cigarette cues would predict risk-taking behavior. Moreover, we investigated these relationships in a sample of both regular smokers and tobacco chippers, guided by the assumption that the relationship between emotion and smoking might differ for smokers with different dependence and smoking frequency rates.