Striatal dopamine mediates the interface between motivational and cognitive control in humans: evidence from genetic imaging

Abstract

Dopamine has been hypothesized to provide the basis for the interaction between motivational and cognitive control. However, there is no evidence for this hypothesis in humans. We fill this gap by using fMRI, a novel behavioral paradigm and a common polymorphism in the DAT1 gene (SLC6A3). Carriers of the 9-repeat (9R) allele of a 40 base pair repeat polymorphism in the 3' untranslated region of DAT1, associated with high striatal dopamine, showed greater activity in the ventromedial striatum during reward anticipation than homozygotes for the 10-repeat allele, replicating previous genetic imaging studies. The crucial novel finding is that 9R carriers also exhibited a greater influence of anticipated reward on switch costs, as well as greater activity in the dorsomedial striatum during task switching in anticipation of high reward relative to low reward. These data establish a crucial role for human striatal dopamine in the modulation of cognitive flexibility by reward anticipation, thus, elucidating the neurochemical mechanism of the interaction between motivation and cognitive control.

Figure 1

Example trials from the experimental paradigm. In both of these trials, the reward-cue indicated that the participant could earn 10 cents with a correct and sufficiently quick response (as opposed to 1 cent in the low-reward condition). The task-cue told the participant to respond to the arrow of the incongruent arrow-word stroop-like target in the first trial, but to the word of the incongruent arrow-word stroop-like target in the second trial. Hence, the second trial is an example of a switch of the task relative to the previous trial. Immediately after the response with a button box, feedback was given with the amount of reward the participant had earned for this specific trial. There was a variable delay of 2–6 s between cues and targets in which participants had to fixate on an asterisk in the middle of the screen.

Figure 2

Effects of dopamine on task switching as a function of reward. (a) Coronal section showing increased left caudate nucleus activity for the three-way interaction on the targets between reward anticipation, task switching, and DAT1 genotype in orange, and the main effect of reward anticipation during reward-cues in red (voxel-level cutoff at P<0.001, uncorrected). L=left; R=right. (b) For illustration purposes, we plotted the switch effect (switch—repeat targets) from the suprathreshold voxels in the dorsomedial striatum. Bars represent the average across subjects; symbols represent individual data points; dotted lines connect the data points belonging to the same subject. All 9R carriers showed larger switch-related activity in the dorsomedial striatum when anticipating high reward relative to the anticipation of low reward. (c) The dopamine-modulated effect of anticipated reward on task switching in the dorsomedial striatum (DMS) correlated significantly with the reward anticipation effect in the ventromedial striatum (VMS) (r=0.49, P=0.03), with the 9R carriers (DAT 9/9) showing both increased activity in dorsomedial striatum as well as ventromedial striatum compared with the 10R homozygotes (DAT 10/10). Note that the plotted data were extracted from the suprathreshold voxels representing the three-way interaction between genotype, reward, and task switching and the main effect of reward anticipation, respectively. (d) The error switch cost (switch—repeat trials). Bars represent the average across subjects; symbols represent individual data points; dotted lines connect the data points belonging to the same subject. Almost all 9R carriers showed a smaller switch cost when anticipating high reward relative to the anticipation of low reward.

Figure 3

Reward anticipation and reward receipt in the striatum. (a) Whole-brain results for the main effects of reward reveal a medial-to-lateral gradient as a function of task phase in ventral striatum: The effect of reward anticipation during reward-cues (in red) was largest in the ventromedial part of the striatum. The effect of reward anticipation during task-cues (in green) is observed more laterally in the striatum. Finally, the effect of reward receipt (in dark blue) was largest in the most lateral parts of the ventral striatum (voxel-level cutoff at P<0.001, uncorrected). L=left; R=right. (b) The reward anticipation effect in the anatomically defined (left) caudate nucleus is plotted for the 10R homozygotes (DAT 10/10) and the 9R carriers (DAT 9/10). The 9R carriers (with presumably more striatal dopamine), but not the 10R homozygotes, showed a significant reward anticipation effect (see ^*) in this caudate ROI, which includes the nucleus accumbens. Error bars represent standard errors of the differences between high- and low-reward-cues. (c) The reward receipt effect in the anatomically defined (left) putamen is plotted for the 10R homozygotes (DAT 10/10) and the 9R carriers (DAT 9/10). The 10R homozygotes (with presumably less striatal dopamine), but not the 9R carriers, showed a significant reward receipt effect (see ^*) in this putamen ROI. Error bars represent standard errors of the differences (SED) between positive feedback and negative feedback.