One of the greatest public health threats on record, smoking is linked to more than 8 million deaths globally each year (WHO, 2020). In the United States, about 55% of adults who smoke make a serious attempt to quit each year (Creamer et al., 2019, Gitchell et al., 2017). Of those, >92% do not maintain abstinence beyond 6 months (Creamer et al., 2019). Chief among reasons for lapses are negative affect (e.g., anxiety, psychological withdrawal) and craving (Sinha, 2011, Zhou et al., 2009, Messer et al., 2018, Piper et al., 2011). With the advent of new intervention tools, including targeted brain stimulation, a better understanding of the neural substrates underlying negative affect and craving during abstinence can help reduce recurrence of smoking.

Clinical neuroscience research has often focused on function of the reward network and the threat network in addiction and mental health broadly (Tabibnia, 2020, Nelson et al., 2013). For example, clinical depression has been characterized with dampened response in the reward network and heightened response in the amygdala and hypothalamic-pituitary-adrenal axis (Charney, 2004), whereas Tobacco Use Disorder (TUD) may involve normal threat response and aberrant reward response (Kunas et al., 2021). Similarly, craving has classically been associated with reward-related circuitry (Volkow and Morales, 2015) and anxiety with threat-related circuitry (Tye et al., 2011).

More recently, the default mode network (DMN) has emerged as another network that is relevant across mental health disorders (Tabibnia, 2020, Whitfield-Gabrieli and Ford, 2012), including TUD and other substance use disorders (Zhang and Volkow, 2019). The DMN is a distributed neural network involved in several functions, notably self-referential thought, such as thinking about one’s past or future (Andrews-Hanna et al., 2014, Buckner and DiNicola, 2019). It tends to be active when a person is not engaged in tasks that demand externally-focused attention – i.e., during a wakeful “resting state” (Buckner and DiNicola, 2019, Raichle et al., 2001). Resting state functional connectivity (RSFC) in the DMN is associated with multiple processes, including negative subjective states (Zhang et al., 2020, Zhou et al., 2020, Lehmann et al., 2016, Li et al., 2020). For example, hyperconnectivity and hyperactivation of the core regions of DMN, particularly posterior cingulate cortex (PCC) and medial prefrontal cortex (MPFC), have been linked with rumination and negative mood (Zhou et al., 2020, Christoff et al., 2016).

The first aim of the current study was to examine the role of the DMN in negative affect and craving during withdrawal from smoking. In people who smoke, RSFC within DMN during abstinence is stronger than after smoking (Zhao et al., 2019, Ding and Lee, 2013, Huang et al., 2014, Janes et al., 2014, Li et al., 2017) and stronger than in people who do not smoke (Vergara et al., 2017, Huang et al., 2014). In particular, early abstinence may increase RSFC in PCC and other DMN regions (Li et al., 2017). However, it is unclear whether RSFC within DMN is related to craving and negative affective states during abstinence. Previous studies have shown positive correlations of RSFC within DMN with cigarette craving (Zhao et al., 2019, Huang et al., 2014, Janes et al., 2014) and withdrawal (Cole et al., 2010). However, these studies had small sample sizes (Cole et al., 2010, Huang et al., 2014, Janes et al., 2014) (n<18) or did not include female participants (Zhao et al., 2019). No study has evaluated the role of DMN RSFC in negative affective states, such as anxiety, during withdrawal from smoking.

The second aim of this study was to examine the similarity of pathways underlying negative affective states and craving. Craving and negative affect are often associated with one another (Watson et al., 2018, Wolitzky-Taylor and Schiffman, 2019, Fatseas et al., 2018, Giuliani and Berkman, 2015, Knapp et al., 2021) and may both contribute to substance use and its recurrence after quitting (Watson et al., 2018, Fatseas et al., 2018, Killen and Fortmann, 1997, Leventhal and Zvolensky, 2015, Piper et al., 2011); therefore, a similar mechanism may underlie both. Furthermore, anxiety and cigarette smoking tend to be comorbid (Leventhal and Zvolensky, 2015) and may have overlapping psychological mechanisms (Ameringer and Leventhal, 2010, Brewer and Roy, 2021). However, a common neural basis of craving and negative affective states within DMN remains unreported. By moving away from the dichotomy of threat and reward and towards a more unifying approach, demonstrating that a common mechanism underlies both subjective states could help identify therapies that address more than one symptom or that are transdiagnostic.

Using secondary analysis of data from a sample of 46 men and women, we tested the hypotheses that after overnight abstinence from smoking 1) DMN RSFC is associated specifically with craving and negative affective states, and 2) a similar pathway underlies both subjective states, as indicated by correlated self-report measures and partially overlapping DMN connectivity patterns. To mitigate issues related to analytical variability (Botvinik-Nezer et al., 2020), we implemented two different analysis pipelines. Taking a confirmatory approach, we used seed-based analyses to identify correlations between subjective states and RSFC of PCC within DMN. These analyses focused on connectivity within DMN, because within-network RSFC is more predictive of smoking status than between-network RSFC (Pariyadath et al., 2014). The PCC was selected as seed because of its involvement in negative subjective states (Renner et al., 2017, Zhang et al., 2020, Zhou et al., 2020, Berman et al., 2014) and smoking abstinence (Li et al., 2017), and because of its potential as a therapeutic target in both anxiety and substance use disorders (Brewer and Garrison, 2014). In a second, more exploratory approach, we used independent component analysis (ICA) with dual regression in whole-brain analyses to identify brain regions in which DMN RSFC correlates with subjective states.