All Cravings are Not Created Equal

We argued that it is important to consider urge intensity when examining the neural correlates of craving [1]. While we believed there were important conceptual and empirical reasons for this position (illustrated empirically by differences between deprived and nondeprived samples), we noted that researchers often have ignored the potential significance of urge levels. We were gratified to see that the commentators generally seemed comfortable with what we had wondered might be viewed as a controversial critique [2–4]. We hope that moving forward researchers conducting “craving” studies with nondeprived smokers will address whether their findings may instead relate to less intense states of desire.

Moeller et al. [2] suggest that the concept of craving requires refinement and propose additional factors besides deprivation that might distinguish between “low-level desire and clinically-relevant craving.” We agree that deprivation is just one factor addressing the broader issue that urge intensity warrants increased attention [5] and that the role of deprivation in craving varies across drugs of abuse. We also concur that improved assessment of craving is needed in neuroimaging research, although doing so in the context of a multi-trial design may prove problematic due to the likely protracted duration of robust craving responses and potential carryover effects [6, 7].

Potvin found “inspiring” our central point that urge intensity matters, but requested additional data to address potential overlapping activation between studies of deprived and nondeprived smokers, which we provide in supplemental tables. Deprived and nondeprived samples yielded largely divergent activation patterns, with only 3 of 22 foci identified across study types overlapping. Although we observed some overlap in the rACC, the overlapping portion was significantly smaller than the region exhibiting more reliable activation for deprived versus nondeprived studies (240 mm² versus 960 mm², respectively). We hope that these new data, coupled with the observation that cue exposure generates more reliable activation across a much larger portion of the rACC in deprived groups provides more persuasive evidence of our general premise that uncontrollable and moderate cravings may trigger different brain responses, especially in light of recent work linking the rACC to clinically-relevant reductions in craving [8–13].

We agree with Thayer and Weiland [4] that there still may be value to studying smokers experiencing mild desire. We believe, though, that study designs should be guided by theory and the specific goals of the study. Like Thayer and Weiland, we believe that our understanding of addiction will benefit not only from studying brain responses during states of intense desire (craving), but also by examining brain activity when this desire is satisfied [e.g., 14] or actively inhibited [e.g., 15]. As they note, addressing the potential influence of acute nicotine (and nicotine withdrawal for that matter) on hemodynamics and neurovascular coupling is also warranted, although effects may be complex [e.g., 16]

In sum, we agree with all the commentators that there may well be a host of variables that influence the neural correlates of craving. We hope that this set of papers and commentaries reinforces the view that not all cravings are equivalent, that there is a need for increased focus on the manner in which craving is conceptualized, manipulated and measured, and that deprivation level is a particularly important variable to consider.