Reward and Affective Regulation in Depression-Prone Smokers

Abstract

Background

There is a disproportionately high smoking prevalence among individuals who are prone to depression. While depression has been conceptualized as a disorder of dysregulated positive affect and disrupted reward processing, little research has been conducted to determine the role of smoking in these processes among depression-prone smokers.

Methods

Depression-prone smokers (DP+; n = 34) and smokers not depression-prone (DP-; n=49) underwent two laboratory sessions, once while smoking abstinent and once while smoking ad-libitum, to assess the relative reinforcing value of smoking and reward sensitivity. Using experience sampling methods, participants completed self-report measures of subjective reward, positive affect, and negative affect across three days while smoking as usual and three days while smoking abstinent.

Results

DP+ were two times more likely to work for cigarette puffs versus money in a progressive ratio, choice task (OR 2.05; CI 95% 1.04 to 4.06, p=0.039) compared to DP-. Reward sensitivity as measured by the signal detection task did not yield any significant findings. Mixed models regressions revealed a 3-way interaction (depression group, smoking phase, and time) for subjective reward, negative affect and positive affect. For all three of these outcomes, the slopes for DP- and DP+ differed significantly from each other (p's < 0.05), and the effect of smoking (vs. abstinence) over time was greater for DP+ than DP- smokers (p's <0.05).

Conclusions

These findings indicate that the effects of smoking on reward and positive affect regulation are specific to DP+ smokers and highlight novel targets for smoking cessation treatment in this population.

Introduction

About 30-60% of smokers seeking to enroll in smoking cessation programs have had at least one lifetime episode of Major Depressive Disorder (MDD) (1-5), compared to 15% of the general population (6). Further, half of smokers who enroll in smoking cessation programs have elevated depression symptoms (2, 4, 7-12). Smokers with a past history of depression and current symptoms of depression (i.e., depression-prone; DP+ smokers) tend to have more difficulty quitting and are at significantly higher risk of relapse than smokers not prone to depression (DP-) (7, 13, 14).

Negative affect has long been considered a critical factor in smoking maintenance among smokers in general, and depression-prone smokers specifically (15, 16). To date, smoking cessation interventions for DP+ smokers have focused on the management of negative affect through behavioral and anti-depressant therapy with little success (5, 17-24). Evidence indicates that targeting negative affect does not significantly improve smoking cessation rates or mitigate negative affect (17, 21, 23, 24), but can exacerbate negative affect and depression symptoms, and decrease the likelihood of quitting smoking (25).

While depression has been conceptualized as a disorder of dysregulated positive affect and disrupted reward processing (26), we know little about the role of smoking in these processes among DP+ smokers (24, 27). Behavioral Economic Theory (BET) provides a framework to integrate features of depression and the effects of smoking. According to BET the choice of one rewarding behavior (e.g., smoking) depends on the presence of alternative reinforcers and the reinforcing value of a drug can be enhanced, or reduced, based on the presence of alternative(s) (28-36). Of relevance, individuals prone to depression have fewer alternative reinforcers (37-39), derive less reward from natural reinforcers in their environment (26, 40-45), and as a result, have diminished positive affect (45-48).

In the context of fewer alternative reinforcers, smoking may be an easily available reinforcer with heightened reinforcing value for DP+ smokers (34, 49). DP+ smokers are two times more likely to rate smoking as their preferred activity (49), have greater smoking-induced dopamine release (50), and to find smoking more reinforcing than DP- smokers (51). Nicotine may also enhance the reinforcement derived from available alternative reinforcers. Pre-clinical models suggest that nicotine potentiates reward from available reinforcers by increasing the sensitivity of brain reward systems or the ability to experience pleasure (27, 52, 53), while nicotine withdrawal decreases reward sensitivity (54). In the context of fewer alternative reinforcers, smoking may be a critical reinforcer, while also increasing pleasure derived from available alternative reinforcers.

Research also indicates that nicotine increases positive affect (55-57), and smoking abstinence decreases positive affect, which predicts smoking relapse (24, 58-60). Nicotine's effect on positive affect may be especially important for DP+ smokers. Smokers with a history of MDD showed heightened positive affect to a positive mood induction when smoking (27). Although largely ignored, smoking may help regulate positive affect across time, rather than simply managing negative affect for DP+ smokers (48, 51, 61).

Supported by converging pre-clinical and clinical evidence, this study sought to provide initial evidence as to whether DP+ smokers compared to DP- smokers: (1) find smoking more reinforcing relative to other reinforcers; and (2) have greater changes in reward sensitivity, subjective reward from self-selected alternative reinforcers, and affect while smoking as usual versus abstinent. A better understanding of smoking's role among DP+ smokers may shed light on novel smoking cessation treatment targets for these smokers.

Methods and Materials

Study Participants

Cigarette smokers (n=83) were recruited from the community through print advertisements. Interested smokers completed a telephone screen assessing smoking history, depression status, medical and psychiatric conditions, and medication/drug use. Eligible smokers were between the ages of 18-65, who smoked at least 10 cigarettes a day for 6 months and could be classified into one of two depression status groups. DP- included smokers with no past history of major depression and no current depression symptoms. DP+ included smokers with both a past history of major depression and current depression symptoms; a group with disproportionate smoking burden and less success at quitting smoking (14, 62). Smokers who had a past history of MDD, but no current depression symptoms were excluded as were smokers with no past history of MDD with current depression.

Past history of major depression was determined by the Inventory to Diagnose Depression-Lifetime (IDD) (51, 63, 64). The IDD is a 22-item scale with response options ranging in symptom severity (0-4) and an additional question regarding the persistence of the symptom for > 2 weeks, resulting in a positive or negative history of depression. Current depressive symptoms were defined as a score > 16 on the 20-item CESD, which correlates with clinical ratings of depression and has high internal consistency (17, 65, 66). Both the IDD and the CESD were completed through a semi-structured telephone interview by masters-level trained psychologists to prevent ineligible participants from attending an intake visit.

Exclusion criteria included: pregnancy, lactation, chronic medical condition, current diagnosis or history of bipolar disorder, schizophrenia or substance abuse (other than nicotine) (27, 51, 60, 67), current or recent use of smoking cessation medications, antidepressant or antipsychotic medications. The exclusion criteria were assessed via self-report on the telephone screen with an objective assessment of smoking status, medication and drug use, and pregnancy at the intake visit. Participants provided written informed consent to a protocol approved by the University of Pennsylvania's Institutional Review Board. Participants then provided a carbon monoxide (CO) breath sample to verify smoking status and a urine sample for a drug screen (Instant Technologies, Inc. Norfolk, VA) and pregnancy. Participants who had a CO < 10 ppm or a positive urine drug screen for illicit drugs or psychotropic medication were excluded. The final sample consisted of 34 DP+ smokers and 49 DP- smokers. Table 1 characterizes the groups.

Table 1.

Descriptive characteristics of DP- (n=49) and DP+ (n=34) smokers at baseline.

	DP -		DP+		Overall		P-value χ21df;t-test
Covariate	Mean	SD	Mean	SD	Mean	SD
Age	41.02	14.17	34.47	13.86	38.34	14.33	0.04
Smoking Rate (cigs/day)	20.10	21.60	18.06	6.99	19.29	17.32	0.60
Nicotine Dependence (FTND)	5.61	1.79	5.85	1.58	5.71	1.70	0.50
MNWS (Baseline)	8.99	5.56	21.59	7.86	13.98	9.00	<0.0001
Positive Affect (PANAS)	35.10	7.29	27.85	7.51	32.13	8.16	<0.001
Negative Affect (PANAS)	13.57	3.75	25.11	8.75	18.30	8.47	<0.001
Complementary Reinforcers	72.06	41.52	65.59	34.04	69.41	38.54	0.46
Substitute Reinforcers	31.65	23.43	20.09	19.53	22.00	31.84	0.02
IDD	5.83	7.96	40.42	7.12	19.52	18.63	<0.0001
CESD	7.14	5.33	30.61	10.66	16.43	13.94	<0.0001
	%	Count	%	Count	%	Count
Female	27%	(13)	47%	(16)	35%	(29)	0.06
Race (White)	37%	(18)	44%	(15)	40%	(14)	0.50
Attended College	51%	(25)	53%	(18)	52%	(18)	0.90
Married	12%	(6)	18%	(6)	14%	(5)	0.50
Employed	53%	(26)	68%	(23)	59%	(20)	0.18

Procedure

Eligible participants completed a baseline assessment of demographics, smoking history and mood and were scheduled for two morning laboratory assessment sessions. One was scheduled after overnight smoking abstinence (9 hours) and the second was scheduled while smoking as usual. The relative reinforcing value of smoking was assessed after smoking abstinence (to ensure motivation to respond) and reward sensitivity was assessed during both smoking abstinence and smoking ad-libitum. Participants received $30 compensation for each laboratory visit. Using experience sampling methods, participants then completed measures of subjective reward, positive affect, and negative affect via telephone across three days while smoking as usual and three days while smoking abstinent. Detailed procedures are described below and depicted in Figure 1.

Figure 1.

Flow chart of study visits and assessments.

Laboratory Assessments

A validated smoking choice paradigm permitted the evaluation of the relative reinforcing value of smoking (RRVS), which is the preference for smoking over other alternatives (67-70). In this paradigm, participants moved a computer mouse to hit targets on one of two computer screens: one to earn points toward smoking and one for money. Using a concurrent schedule (71, 72), participants could switch from working on one screen to the other as often as they wish. Participants were instructed to move the computer mouse to have the cursor hit the targets (either a $ or a cigarette). Consistent with relative reinforcement paradigms, the reinforcement schedule in the money-earning screen remained constant at a fixed ratio FR-25 (25 targets achieved to earn a point) while the reinforcement schedule for smoking increased (require more effort) with a progressive ratio schedule of PR-25x over 10 trials, such that 25, 50, 75, 100, 125, 150, 175, 200, 225, and 250 targets had to be achieved to earn a point (73, 74). The task was performed for 10 trials for a total of 10 points from which they earned either $0.25 for each point (i.e., up to $2.50 paid at the end of the procedure) or one puff of a cigarette for each point (i.e., up to 10 puffs, smoked at the end of the procedure to prevent satiation from influencing susequent responding), or any combination of both reinforcers. Participants then began a 1-hour wait in the lab to standardize session duration and ensure that responding was based on reinforcer preference.

A signal detection task was used to measure reward sensitivity. The participants’ goal was to determine, via key press, whether a short (11.5mm) or a long (13mm) mouth was presented on a previously mouth-less cartoon face for 100 ms (42, 75, 76). During each of the two 100-trial blocks, reward feedback (“Correct!! You win 5 cents”) was presented after 40 correct trials according to a controlled reinforcement schedule. One stimulus type (either the short or the long mouth) was randomly selected to be disproportionately rewarded (labeled the “rich stimulus”; 30 trials) for correct responses at a 3:1 ratio. The less-frequently rewarded stimulus was labeled the “lean stimulus” (10 trials). Participants were not informed of this reward disparity. Reward feedback for correct responses was given according to a pseudo-randomized schedule that determined which of the 10 lean stimulus and 30 rich stimulus trials of the 100 total trials were rewarded for correct identifications within a block. Reward feedback was displayed for 1500 ms and followed by a blank screen for 250 ms. For the entire task, each participant “earned” around $6 that was paid at the end of the session.

Experience Sampling

After the second laboratory session, participants received instructions for either a 3-day smoking abstinence phase or a 3-day smoking ad-lib phase. Each phase involved Experience Sampling, where participants were called via a study cell phone at 7 randomly determined times during a 14-hour day (e.g., 8am - 10pm) for a 1-2 minute, 12 question interview. Each interview assessed positive affect, negative affect, current activity, and the level of pleasure derived from the activity (subjective reward). The specific interview items are described under Subjective Outcomes, below.

The times of the 7 random assessments were computer generated and distributed evenly throughout the day (i.e., stratified block sampling) for each subject for each day of each phase. To promote compliance, participants received $5 for each phone call completed (a total of 21 calls for each phase) and a telephone number to call with issues (77, 78). The order of the abstinent and ad-lib phases was randomly determined for the depression status groups, and counterbalanced to minimize bias due to order effects. Both phases included one weekend and two week days of assessment, with over four days between phases. Consistent with the literature (78-80), compliance with the telephone calls was >80%. Participants who were unable to comply with study procedures were withdrawn from the study.

In preparation for the abstinence phase, participants received general instructions and a 20-minute coaching session on tips for remaining abstinent (81, 82). Participants refrained from smoking beginning 12:01 am on Day 1 of the 3-day Abstinence phase. During the 3-day abstinence phase, abstinence was verified daily via CO < 5 ppm and a Nic Alert saliva cotinine ≤ 1 (Nymox Pharmaceutical, Hackensack, NJ) (81, 83, 84). The NicAlert salivary cotinine test was performed on days 2 and 3, but not day 1 due to residual elevations from prior smoking. Participants received $100 for remaining smoke-free across the 3-day abstinence phase (81, 82, 84). Two DP+ and two DP- participants were unable to remain smoking abstinent and were withdrawn from the study.

Objective Outcomes

Relative reinforcing value of smoking was defined as the maximum amount of responding for smoking versus money across the 10 trials (73, 74). For the reward sensitivity task, response bias was calculated by a formula derived from a behavioral model of signal detection (85). Change in response bias from block 1 to 2 was used as the index of reward sensitivity (76). Outlying trials were removed following Bodgon and Pizzagalli (75).

Subjective Outcomes

Experience Sampling via brief phone interviews enabled us to capture prospective, random samples of smoking and abstinence related changes in affect and subjective reward from activities across smoking and abstinence days in the participants own environment in “real-time” avoiding retrospective recall (86, 87). Consistent with previous research, participants rated their positive affect (4 items: happy, relaxed, content, enthusiastic) and negative affect (5 items: sad, stressed, irritable, bored, tense) “at that moment” on a scale from 0=not at all to 4=extremely (45, 79, 88-90). Subjective reward from self-selected reinforcers was measured by asking participants to indicate what activity they were currently engaged in and how much they were enjoying that activity on a 5 point scale from 0=not at all to 4=very much (45, 91). These types of ecological momentary approaches produce valid and reliable assessments of affect and reward in substance using populations (45, 88-90, 92, 93).

Covariates

Baseline affect, nicotine dependence, and alternative reinforcers were measured to characterize the sample and to potentially control for their effects in the multivariate models. The Fagerstrom Test for Nicotine Dependence (FTND) is a 6-item measure with good internal consistency (α= .64) and high test-retest reliability (r = .88) (94). The Minnesota Nicotine Withdrawal Scale was used to measure withdrawal symptoms (95). The Positive and Negative Affect Schedule (PANAS) was used to measure positive and negative affect “over the past week” (1=not at all to 5= extremely (96) via two, 10-item scales with excellent psychometric properties (24, 27, 51, 96). Alternative Reinforcers were measured by the adapted Pleasant Events Schedule (PES) (39), which was designed to assess reinforcers that occur in an individual's natural environment. The 78-items were rated once in terms of frequency (0=none to 2=often) and once in terms of enjoyability (0=none to 2=very) over the past 30 days, yielding a frequency score, an enjoyability score, and the cross product is the reinforcement from the activity (39, 97-100).

Data Analysis

All analysis was conducted using Stata 13.0. Univariate statistics were generated to describe the study population. Preliminary bivariate analyses used contingency table methods (χ²) or t-tests as appropriate. Multivariate models used the Stata XT series of routines for population averaged repeated measures analysis (GEE), including mixed-effects regression for continuous measures, and mixed-effects logistic regression for binary outcomes.

Results

Descriptive Statistics

Table 1 presents the frequency distributions for all categorical variables, and the means and standard deviations for the continuous model variables. Consistent with prior work (14, 101-103), DP+ subjects were younger (35 years versus 41) and more were female (47% vs. 27%). Groups did not differ significantly on other characteristics, except for those associated with depression status (e.g., baseline affect, substitute alternative reinforcers). Table 2 provides univariate statistics for outcome variables by group and phase. Gender and age were considered in each model. Because they accounted for less than 1.5% combined variance in each model, they were not retained. The average effect sizes for the interactions described below are as follows: RRVS .60; positive affect .32; negative affect .66; and subjective reward .13. Reward sensitivity was an exploratory variable with an observed effect size of .10.

Table 2.

Univariate statistics for outcome variables by group and phase.

	DP-	DP+	DP-	DP+
Outcome	Smoking		Abstinent
	Mean SD	Mean SD	Mean SD	Mean SD
RRVS	1.03 (1.43)	1.90 (2.66)
Response Bias	0.13 (0.24)	0.10 (0.15)	0.12 (0.16)	0.14 (0.16)
Positive Affect Day 1	12.52 (3.77)	9.48 (3.39)	12.15 (3.84)	10.16 (3.62)
Positive Affect Day 2	12.32 (3.86)	9.27 (3.42)	12.17 (4.20)	10.02 (3.64)
Positive Affect Day 3	12.38 (3.93)	10.09 (3.74)	12.10 (3.67)	10.04 (3.82)
Negative Affect Day 1	7.41 (2.86)	12.44 (5.77)	7.08 (3.13)	10.39 (5.08)
Negative Affect Day 2	7.18 (2.87)	12.47 (5.68)	7.43 (3.54)	10.90 (5.31)
Negative Affect Day 3	6.91 (2.72)	11.46 (5.78)	7.29 (3.26)	10.85 (5.37)
Subjective Reward Day 1	2.63 (1.23)	1.85 (1.26)	2.49 (1.15)	1.80 (1.18)
Subjective Reward Day 2	2.40 (1.31)	1.80 (1.27)	2.47 (1.22)	1.87 (1.26)
Subjective Reward Day 3	2.55 (1.24)	2.10 (1.28)	2.51 (1.15)	1.93 (1.18)

Multivariate Models

Relative Reinforcing Value of Smoking

The choice to work for cigarette puffs versus money was modeled using logistic regression, with effort (i.e., number of responses) treated as a continuous variable, and depression group treated as a binary variable. We controlled for alternative reinforcers. As the effort required to earn a cigarette puff increased, the probability of working for cigarette puffs decreased (OR 0.86; CI 95% = 0.80 to 0.92), p<0.001). However, smokers in the DP + group were two times more likely to work for cigarette puffs than smokers in the DP- group, despite increasing levels of effort (OR 2.05; CI 95% = 1.04 to 4.06, p=0.039). Figure 2 depicts these findings.

Figure 2.

The probability of choosing smoking versus money as effort required to earn cigarette puff increased.

Reward Sensitivity

We tested reward sensitivity as measured by the signal detection task by modeling response bias using mixed models regression for the paired data. Neither depression status group nor smoking phase was predictive of response bias (p's > 0.24).

Subjective Reward, Positive Affect, and Negative Affect

The subjective reward derived from self-selected activities in the participants’ natural environment, positive affect and negative affect were modeled using mixed models regressions (Table 3; Figure 3). The models controlled for telephone call within day, and the difference in withdrawal scores between abstinent and smoking phases. This is important as the DP+ group's withdrawal symptoms were twice that of DP- when smoking (13.64 vs 6.64) and when abstinent (20.03 vs 9.72). Order of the smoking and abstinent phase was tested and found to be nonsignificant.

Table 3.

Mixed models regression of subiective reward, negative affect, and positive affect.

Subiective Reward
Predictor	Coefficient (C195%)	P-Value
DP - AND Abstinent	0-Reference Group
DP - AND Smoking	0.14 (−0.12 to 0.39)	p=0.284
DP + AND Abstinent	−0.37 (−0.74 to −0.01)	p=0.046	Heterogeneity of slopes:
DP + AND Smoking	−0.79 (−1.15 to −0.42)	p=0.000	χ2(3)=9.37, p=0.025
Day by DP - AND Abstinent β1	0.02 (−0.06 to 0.10)	p=0.620
Day by DP - AND Smoking β2	−0.03 (−0.12 to 0.05)	p=0.446	Contrast (β3-β4)-(β1-β2):
Day by DP + AND Abstinent β3	−0.04 (−0.14 to 0.06)	p=0.472	χ2(1)=6.59, p=0.01
Day by DP + AND Smoking β4	0.15 (0.05 to 0.25)	p=0.003
Negative Affect
DP - AND Abstinent	0-Reference Group
DP - AND Smoking	0.61 (−0.09 to 1.30)	p=0.087
DP + AND Abstinent	2.73 (1.18 to 4.28)	p=0.001	Heterogeneity of slopes:
DP + AND Smoking	6.08 (4.53 to 7.63)	p=0.000	χ2(3)=25.47, p<0.001
Day by DP - AND Abstinent β1	0.06 (−0.17 to 0.29)	p=0.601
Day by DP - AND Smoking β2	−0.29 (−0.52 to −0.06)	p=0.013	Contrast (β3-β4)-(β1-β2):
Day by DP + AND Abstinent β3	0.30 (0.03 to 0.57)	p=0.033	χ2(1)=4.470, p=0.035
Day by DP + AND Smoking β4	−0.60 (−0.87 to −0.33)	p=0.000
Positive Affect
DP - AND Abstinent	0-Reference Group
DP - AND Smoking	0.59 (0.01 to 1.16)	p=0.045
DP + AND Abstinent	−1.39 (−2.77 to −0.01)	p=0.048	Heterogeneity of slopes:
DP + AND Smoking	−2.76 (−4.14 to −1.38)	p=0.000	χ2(3)=10.67, p=0.014
Day by DP - AND Abstinent β1	0.05 (−0.13 to 0.24)	p=0.578
Day by DP - AND Smoking β2	−0.12 (−0.31 to 0.07)	p=0.207	Contrast (β3-β4)-(β1-β2):
Day by DP + AND Abstinent β3	−0.10 (−0.33 to 0.12)	p=0.376	χ2(1)=8.20, p=0.004
Day by DP + AND Smoking β4	0.33 (0.10 to 0.55)	p=0.004

Figure 3.

Slopes reflecting change in subjective reward and affect across time and smoking or abstinent phase. Error bars reflect difference scores for day 3 minus day 1.

Depression group, smoking phase, and day were involved in a 3-way interaction for subjective reward. The interaction is best interpreted as a difference among slopes for the effect of day. For smokers in the DP- group, there was no significant change in subjective reward during the abstinence phase (p=0.62) or during the smoking phase (p=0.45). Although, subjective reward did not significantly decrease for DP+ during the abstinence phase (p=0.47), there was a significant increase during the smoking phase (p=0.003). The slopes for the depression groups differed significantly from each other (χ²(3)=9.37, p=0.025), and the effect of smoking versus abstinence over time was greater for DP+ than DP- (χ²(1)=6.59, p=0.01).

Similar to results for subjective reward, depression group, smoking phase, and day were involved in a 3-way interaction for positive affect as well as for negative affect. Positive affect did not change significantly for DP- during the abstinence phase (p=0.58) or the smoking phase (p=0.21). Positive affect did not change significantly for DP+ during the abstinence phase (p=0.40), but increased significantly during the smoking phase (p=0.004). The slopes for the two groups differed significantly from each other (χ²(3) =10.67, p=0.014), and the effect of smoking versus abstinence over time was greater for DP+ compared to DP- (χ²(1) =8.20, p=0.004).

For both groups, negative affect decreased during the smoking phase (DP-, p=0.013, and DP+, p<0.001). However, during the abstinence phase there was a significant increase for DP+ (p=0.03), but not for DP- (p=0.60). The slopes for the two groups differed significantly from each other (χ²(3)=25.47, p<0.001), and the effect of smoking versus abstinence over time was greater for DP+ than DP- (χ²(1)=4.470, p=0.035).

Discussion

The present study is the first to show that the relative reinforcing effects of smoking and the impact of smoking on reward and positive affect regulation are specific to DP+ smokers. Compared to DP- smokers, DP+ smokers have a heightened relative reinforcing value of smoking, and enhanced subjective reward from typical activities, and greater positive affect while smoking compared to abstinence. DP- smokers have little affective and reward dysregulation when abstinent, although both DP+ and DP- smokers show decreased negative affect while smoking. These findings suggest that DP+ smokers may be less able to sustain smoking abstinence because they not only lose smoking as an important reinforcer, but they also lose the enhanced positive affect, mitigated negative affect and pleasure from other reinforcers that accompany smoking. As such, the present study uncovers unique vulnerabilities to achieving and maintaining smoking abstinence and highlights novel targets for smoking cessation treatment in this population.

Previous research suggests that smokers with a history of MDD find smoking more reinforcing than smokers without a history of MDD (51). We extend these findings and show that DP+ smokers find smoking two times more reinforcing than other reinforcers compared to DP-smokers. The heightened relative reinforcing value of smoking for DP+ smokers highlights a significant reinforcer deficit upon quitting smoking that occurs within a context of already diminished reinforcers. Preliminary research suggests that behavioral activation smoking cessation treatment may allow DP+ smokers to overcome the loss of smoking as a reinforcer by promoting increased engagement in rewarding activities (104). However, behavioral activation alone may be insufficient to promote long-term smoking cessation among DP+ smokers as it does not appear to impact the enjoyment derived from typically rewarding activities in this population.

Indeed, the present findings highlight that DP+ smokers experience greater enjoyment from their daily activities while smoking ad libitum than while abstinent. For DP- smokers, subjective reward changed little regardless of smoking status. These differences in reward responsivity cannot be attributed to between-group differences in nicotine withdrawal symptoms. If smoking permits DP+ smokers to derive greater reward from typical reinforcers, smoking cessation treatment that includes behavioral skills to magnify or savor the enjoyment derived from daily reinforcers may help ameliorate the pleasure deficit (105, 106). These behavioral strategies derived from Positive Psychology focus on re-experiencing and dismantling the pleasurable aspects of daily events and have been shown to enhance positive mood in depression-prone populations (106). This is important because DP+ smokers have fewer reinforcers that are not linked to smoking. Recent studies in animals and humans suggest that bupropion and varenicline have reinforcement enhancing effects (107, 108), but surprisingly have not been fully evaluated in DP+ smokers (17, 20).

Evidence indicates that targeting negative mood does not significantly improve smoking cessation rates or mitigate negative mood for DP+ smokers (17, 21, 23, 24). Our results suggest that affective changes from smoking to abstinence are greater for DP+ smokers and that smoking helps to regulate positive affect, rather than simply reducing negative affect. Cognitive-behavioral treatment to lessen negative mood does not alter underlying reward-related substrates among persons with depression (109), which may offer a neurobiological account for its lack of effectiveness for smoking cessation among DP+ smokers. In contrast, behavioral therapy for depression that targets involvement in non-drug rewarding activities to promote positive mood increases neural responses to rewards (110). While there is debate as to whether positive and negative affect are distinct constructs or exist on opposite ends of an affective continuum, current research is focused on characterizing these dimensions to ultimately inform treatment development (111). For instance, targeting positive affect in DP+ smokers may have the added benefit of decreasing negative affect (106). Nevertheless, the present findings support a shift in smoking treatment for DP+ smokers from suppressing negative affect and smoking behavior to enriching positive affect and pleasure derived from alternative reinforcers such that smoking is a less attractive and reinforcing option (61).

Using a probabilistic reward task, we anticipated that the DP+ group would show a decreased response bias toward the reinforcing stimulus when abstinent, but not when smoking. This would have indicated reduced reward responsiveness while abstinent. Neither group of smokers developed a response bias. To detect a response bias in a community sample with elevated depression symptoms, we may have needed a greater number of learning trials versus the shortened version that was administered (76), more than overnight smoking abstinence, or a larger sample size to detect smaller effects. Although speculative, nicotine may help normalize disrupted reward learning that characterizes depression (112-114).

The present study focused on a population with disproportionately high smoking prevalence. Building on a strong theoretical and empirical foundation, its rigorous and ecologically valid experimental procedures included the first evaluation of (1) the relative reinforcing value of smoking, and (2) changes in affect and subjective reward in “real-time” in DP+ smokers’ natural environment. The findings provide initial support for novel smoking cessation treatment targets for a population that is often excluded from smoking cessation clinical trials. Despite these strengths, the limitations of the study should be acknowledged. While the CESD is a reliable and well-established self-report measure of current depression symptoms, we did not determine whether the DP+ participants met criteria for current MDD. Consequently, we cannot be sure that the present findings extend to smokers with past and present MDD. We also are unable to determine if similar relationships between affect and subjective reward exist for smokers who only have a past history of MDD or a current elevation in depression symptoms. Finally, participants were not seeking smoking cessation treatment, thus the findings may not fully generalize to those DP+ smokers who are motivated to quit smoking.

The present study advances smoking research in co-morbid populations by integrating unique deficits in depression with smoking's role in deficit amelioration. DP+ smokers experience a loss of smoking reinforcement, enhanced pleasure from other reinforcers, and positive mood upon smoking abstinence as well as increased negative affect. A smoking lapse could acutely restore these affective and reward-related functioning, which may then drive smoking relapse. This may help explain the difficulty sustaining smoking abstinence and the disproportionate smoking prevalence in this population. Novel behavioral smoking cessation interventions that consider the unique mechanisms that maintain smoking behavior in DP+ smokers may increase quitting success.