Cognitive deficits during nicotine withdrawal may contribute to smoking relapse. However, interacting effects of chronic nicotine dependence and acute nicotine withdrawal on cognitive control are poorly understood. Here, we examine the effects of nicotine dependence (trait; smokers versus non-smoking controls), and acute nicotinic stimulation (state; administration of nicotine and varenicline, two FDAapproved smoking cessation aids, during abstinence), on two well-established tests of cognitive control, the Go-Nogo task and the Flanker task, during fMRI scanning. We compared performance and neural responses between these four pharmacological manipulations in a double-blind, placebo-controlled crossover design. As expected, performance in both tasks was modulated by nicotine dependence, abstinence and pharmacological manipulation. However, effects were driven entirely by conditions that required less cognitive control. When demand for cognitive control was high, abstinent smokers showed no deficits. By contrast, acutely abstinent smokers showed performance deficits in easier conditions and missed more trials. Go-Nogo fMRI results showed decreased inhibition-related neural activity in right anterior insula and right putamen in smokers and decreased dorsal anterior cingulate cortex activity on nicotine across groups. No effects were found on inhibition-related activity during the Flanker task, or on error-related activity in either task. Given robust nicotinic effects on physiology and behavioral deficits in attention, we are confident that pharmacological manipulations were effective. Thus, findings fit a recent proposal that abstinent smokers show decreased ability to divert cognitive resources at low or intermediate cognitive demand, while performance at high cognitive demand remains relatively unaffected, suggesting a primary attentional deficit during acute abstinence.