Greater BOLD activity but more efficient connectivity is associated with better cognitive performance within a sample of nicotine-deprived smokers

Abstract

The first few days of an attempt to quit smoking are marked by impairments in cognitive domains, such as working memory and attention. These cognitive impairments have been linked to increased risk for relapse. Little is known about individual differences in the cognitive impairments that accompany deprivation or the neural processing reflected in those differences. In order to address this knowledge gap, we collected functional magnetic resonance imaging (fMRI) data from 118 nicotine-deprived smokers while they performed a verbal n-back task. We predicted better performance would be associated with more efficient patterns of brain activation and effective connectivity. Results indicated that performance was positively related to load-related activation in the left dorsolateral prefrontal cortex and the left lateral premotor cortex. Additionally, effective connectivity patterns differed as a function of performance, with more accurate participants having simpler, more parsimonious network models than did worse participants. Cognitive efficiency is typically thought of as less neural activation for equal or superior behavioral performance. Taken together, findings suggest cognitive efficiency should not be viewed solely in terms of amount of activation but that both the magnitude of activation within and degree of covariation between task-critical structures must be considered. This research highlights the benefit of combining traditional fMRI analysis with newer methods for modeling brain connectivity. These results suggest a possible role for indices of network functioning in assessing relapse risk in quitting smokers as well as offer potentially useful targets for novel intervention strategies.

Keywords: Addiction; connectivity; fMRI; individual differences; smoking; working memory.

Figure 1

Voxels exhibiting a significant main effect of memory load (3-back vs. 0-back). Regions of voxels are numbered as follows: 1=anterior cingulate cortex/supplementary motor area; 2=left dorsolateral prefrontal cortex; 3=right dorsolateral prefrontal cortex; 4=left lateral premotor cortex; 5=right lateral premotor cortex; 6=left inferior parietal lobule; 7=right inferior parietal lobule.

Figure 2

Schematic of group model and subgroup models based on d-prime quartile. Q4=top performing quartile; Q3=upper middle quartile; Q2=lower middle quartile; Q1=bottom quartile. DLPFC=dorsolateral prefrontal cortex; ACC=anterior cingulate cortex/supplementary motor area; lPMC=lateral premotor cortex; iPar=inferior parietal lobule.