1. CYP2A6 genetic variation shapes striatal-cingulate brain circuits and executive processing in smokers: Variation in the CYP2A6 gene alters the rate of nicotine metabolic inactivation and is associated with smoking behaviors and cessation success rates. The underlying neurobiological mechanisms of this genetic influence are unknown. Intrinsic functional connectivity strength, a whole-brain, data-driven, graph theory-based method, was applied to resting-state functional magnetic resonance imaging data in 66 smokers and 92 nonsmokers. A subset of subjects (n = 23/20; smokers/nonsmokers) performed the monetary incentive delay task, probing reward anticipation, and a go/no-go task, probing response inhibition, on two occasions, in the presence and absence of a nicotine patch. A significant CYP2A6 genotype smoking effect was found in the dorsal anterior cingulate cortex and ventral striatum, such that the normal (vs. slow) genotype individuals showed greater functional connectivity strength among smokers but not nonsmokers. Functional connectivity strength was negatively associated with severity of nicotine dependence in slow metabolizers. Both hubs were biased by inputs from the insula identified from seed-based connectivity. Similar gene environment interactions were seen in ventral striatum during smoking abstinence when subjects performed the monetary incentive delay task and in dorsal anterior cingulate cortex when they performed the go/no-go task; both reductions were normalized in smokers (and increased in nonsmokers) after acute nicotine administration. As the CYP2A6 effect was seen only in smokers, these data suggest that the rate of nicotine metabolism-and thus the concentration of nicotine presented to the brain over the course of nicotine addiction-shapes brain circuits that, among other functions, compute reward and impulsivity processes. 2. Combining multiple resting-state fMRI features during classification: optimized frameworks and their application to nicotine addiction. Withdrawal from nicotine is an important contributor to smoking relapse. Understanding how reward-based decision making is affected by abstinence and by pharmacotherapies such as nicotine replacement therapy and varenicline tartrate may aid cessation treatment. The goal was to independently assess the effects of nicotine dependence and stimulation of the nicotinic acetylcholine receptor on the ability to interpret valence information (reward sensitivity) and subsequently alter behavior as reward contingencies change (cognitive flexibility) in a probabilistic reversal learning task. Nicotine-dependent smokers and nonsmokers completed a probabilistic reversal learning task during acquisition of functional magnetic resonance imaging (fMRI) in a 2-drug, double-blind placebo-controlled crossover design conducted from January 21, 2009, to September 29, 2011. Smokers were abstinent from cigarette smoking for 12 hours for all sessions. In a fully Latin square fashion, participants in both groups underwent MRI twice while receiving varenicline and twice while receiving a placebo pill, wearing either a nicotine or a placebo patch. Imaging analysis was performed from June 15, 2015, to August 10, 2016. A well-established computational model captured effects of smoking status and administration of nicotine and varenicline on probabilistic reversal learning choice behavior. Neural effects of smoking status, nicotine, and varenicline were tested for on MRI contrasts that captured reward sensitivity and cognitive flexibility. The study included 24 nicotine-dependent smokers (12 women and 12 men; mean SD age, 35.8 9.9 years) and 20 nonsmokers (10 women and 10 men; mean SD age, 30.4 7.2 years). Computational modeling indicated that abstinent smokers were biased toward response shifting and that their decisions were less sensitive to the available evidence, suggesting increased impulsivity during withdrawal. These behavioral impairments were mitigated with nicotine and varenicline. Similarly, decreased mesocorticolimbic activity associated with cognitive flexibility in abstinent smokers was restored to the level of nonsmokers following stimulation of nicotinic acetylcholine receptors (familywise error-corrected P<.05). Conversely, neural signatures of decreased reward sensitivity in smokers (vs nonsmokers; familywise error-corrected P<.05) in the dorsal striatum and anterior cingulate cortex were not mitigated by nicotine or varenicline. In conclusion, there was a double dissociation between the effects of chronic nicotine dependence on neural representations of reward sensitivity and acute effects of stimulation of nicotinic acetylcholine receptors on behavioral and neural signatures of cognitive flexibility in smokers. These chronic and acute pharmacologic effects were observed in overlapping mesocorticolimbic regions, suggesting that available pharmacotherapies may alleviate deficits in the same circuitry for certain mental computations but not for others. 3. Machine learning techniques have been applied to resting-state fMRI data to predict neurological or neuropsychiatric disease states. Existing studies have used either a single type of resting-state feature or a few feature types (<4) in the prediction model. However, resting-state data can be processed in many different ways, yielding different feature types containing complementary and/or novel information, leaving uncertain the most informative features to provide to the classifier. In this study, multiple resting-state features were calculated from two main analytical categories: local measures and network measures. Feature selection was adopted using an optimized grid-search approach selecting top ranked features from statistical tests. We then tested three optimized frameworks: feature combination, kernel combination, and classifier combination, all using the support vector machine as an elementary classifier, to combine these resting-state feature types. When applied to nicotine addiction, with a cohort size of 100 smokers and 100 non-smokers, via a 10-fold cross-validation procedure, the feature combination and the classifier combination achieved an accuracy of 75.5%, while the kernel combination achieved a 73.0% accuracy; all three combination frameworks improved classification performance compared to the single feature type based results (best accuracy 70.5%). This study not only reveals the discriminative power of resting-state data, but also demonstrates the efficiency of combining multiple features from one data phenotype to improve classification performance.
Showing the most recent 10 out of 40 publications
