Mood disorders constitute the greatest disease burden in the developed world. Alcohol use disorders are almost as costly, implicated in 20–40% of hospital admissions and frequently comorbid with mood disorders but having a later age of onset (Li et al. 2004b
). The dominant psychosocial models developed to explain the relationship between anxiety disorders and alcohol use disorders posit that alcohol reduces anxiety and other uncomfortable responses to stress in predisposed individuals. This effect is thought to encourage chronic self-medication because of the temporary negative reinforcement alcohol can provide (Morris et al. 2005
). Inconsistent empirical support for these models has underlined the importance of individual differences in risk for alcoholism (Sher and Levenson 1982
), including gender, family history of alcoholism, and personality characteristics such as high novelty seeking and low harm avoidance (Croissant et al. 2008
; Croissant et al. 2006
Presence of the TaqIA A1 allele interacts with stress to predict cognitive functions in adolescents (Berman and Noble 1997
) and alcoholism in adults (Madrid et al. 2001
). It interacts with negative affect and gender to predict alcoholic personality characteristics (Berman et al. 2003
) and severity of substance abuse in adolescents (Conner et al. 2005
). Furthermore, among hospitalized substance abusers or alcoholics (Nixon and Parsons 1990
), and drug-naive adolescent boys carrying the A1 allele, but not among boys without it (Berman et al. 2002
), the negative correlation between novelty seeking and harm avoidance that generally characterizes personality assessment is reversed. This finding suggests that rather than providing positive reinforcement, novelty seeking may serve a self-medicating negative reinforcement function in both severe substance abusers and in boys with the A1 allele. We predicted that acute ethanol administration would be more reinforcing in A1+ than A1− social drinkers, as indicated by ameliorated anxiety and fatigue, and altered brain activity in structures associated with negative affect.
Previous studies that used blood-sampling during PET to determine absolute metabolic rate for glucose have demonstrated that acute doses of ethanol equivalent to 2–3 mixed drinks produce robust decreases in cerebral glucose metabolism (> 20% in the posterior cerebral cortex) (Volkow et al. 2008
; Volkow et al. 2006b
). of the 2008 paper (Volkow et al. 2008
), also presents relative (whole-brain normalized) metabolic images like those employed in the current study. These images depict a more modest relative decrease in posterior cortical metabolism and a relative increase in subcortical and temporal cortices, particularly in the left hemisphere. Although the current study formally assessed alcohol effects only within regions previously associated with affect or reinforcement, visual examination of results in posterior cortex confirmed the decreased relative metabolism previously reported (Volkow et al. 2008
), just as the relative increases in striatal and amygdala metabolism in are consistent with the prior results in these structures. These observations suggest that the relative increases in and represent relative preservation of glucose metabolism (i.e., less decrease than the global average). Individual values in the areas of interest are consistent with A1+ men but not A1− men having relatively preserved glucose metabolism in the insula and striatum, and with A1− men having a greater than whole brain decrease in glucose metabolism in the anterior cingulate cortex.
Although dopamine release in the nucleus accumbens has been associated with alcohol-related reinforcement (Volkow et al. 2007
), dopamine release in the dorsal striatum, where the current effects are largest, has also been associated with craving for alcohol (Heinz et al. 2005
), and for other drugs (Volkow et al. 2006c
; Wong et al. 2006
). We interpret relative preservation of activity in the striatum and amygdala as indicating affective responses to ethanol, and the greater preservation of activity in the striatum and insula in A1+ men as suggesting larger affective responses. In A1− men, reduction of activity by ethanol in the rostral anterior cingulate, an area associated with cognitive modulation of alcohol feelings (Ingvar et al. 1998
; Sinha and Li 2007
), may indicate fewer alcohol-related feelings, or less ability to modify them.
The lower anxiety and fatigue after ethanol in A1+ men, compared with the higher anxiety and fatigue in A1− men, strongly supports the hypothesis that ethanol is more reinforcing in A1 carriers. The increased negative affect in A1− men is consistent with a recent fMRI study where the dopamine agonist bromocriptine improved speeded motor performance and activation of the nucleus accumbens in A1 carriers, who performed worse than noncarriers under placebo and improved less in a high-incentive condition, but whose deficits were abolished by bromocriptine (Kirsch et al. 2006
). In contrast, bromocriptine lowered performance in A1− individuals, just as ethanol worsened negative affect in our A1− men. The authors suggested that an inverted ‘U’ function may characterize the relationship of dopamine activity to optimal functioning, and that A1 carriers are characterized by both a reward deficiency syndrome and increased delivery of reward by dopamine agonists.
The latter idea is further supported by the current relationship between ethanol-induced change in negative affect and regional brain activity (). Positive covariation of anxiety/fatigue score with thalamic activity across groups is consistent with previous studies of alcohol-induced fatigue (Jia et al. 2007
; Volkow et al. 2006b
). Positive covariation of these scores with ACC activity is consistent with executive-modulation of alcohol feelings (Ingvar et al. 1998
; Sinha and Li 2007
). Negative covariation of concurrent changes in anxiety/fatigue and activity in a cluster comprising 15% of the left insula in A1+ men suggests that ethanol-related reduction in anxiety/fatigue (see ) may reflect the positive interoceptive feelings produced by alcohol, sometimes referred to as a “warm glow.” In A1− men, the more dorsal negative covariation in 7% of left insula is harder to interpret and may be spurious, since it is the only one of 24 comparisons to produce a p < 0.05 cluster (see ).
The dearth of relationships between the effects of ethanol on negative affect and brain activity in A1− men is, however, consistent with a report that nonalcoholic members of alcoholic families exhibit higher D2 receptor availability than alcoholics from the same families (Volkow et al. 2006a
). The study also found associations between D2 availability and glucose metabolism in frontal regions implicated in emotional reactivity and executive control. We speculate that higher levels of D2 receptors in A1− men (Jonsson et al. 1999
; Noble 2003
) could protect against alcoholism by improving regulation of emotional responses that might otherwise lead to drug-seeking.
In contrast to the lack of evidence that changed negative affect was related to change in brain activity in A1− men, the decreased anxiety and fatigue produced by ethanol in A1+ men accompanied decreased activity in the structures most consistently associated with both anxiety (amygdala) (Koob and Le Moal 2008
; Silberman et al. 2008
), and fatigue (thalamus) (Jia et al. 2007
; Volkow et al. 2006b
). Both effects survived a conservative Bonferroni correction. also shows decreases in three of four ROIs comprising a structure previously associated with craving (striatum) (Koob and Le Moal 2008
; Yoder et al. 2008
). In addition, reduced negative affect after ethanol in the A1+ group was accompanied by significant activation of right ventrolateral orbitofrontal cortex. Activity in this region has been associated with lower stress, inhibition of stress-related amygdala activity, and higher psychosocial resources for threat-regulation (Taylor et al. 2008
Limitations of this pilot study include a lack of detailed information on lifetime alcohol consumption, and the restriction of testing to Caucasian males. Although these findings require replication in larger and more diverse samples, they provide the first evidence that the heightened risk for alcoholism associated with the DRD2 A1 allele may be mediated by increased reward sensitivity and negative reinforcement when drinking alcohol.