Dopamine effects on human error processing depend on catechol-O-methyltransferase VAL158MET genotype

Abstract

Brain dopamine (DA) has been linked to error processing. Because high and low (vs medium) prefrontal cortex (PFC) DA levels may facilitate D2-receptor-related modulations of PFC neural activation patterns, we hypothesized that high and low DA predicts increased error-specific transitions of PFC activity. Male human participants (n = 169) were genotyped for the catechol-O-methyltransferase (COMT) Val158Met polymorphism, associated with low (Val) and medium (Met) PFC DA levels. In addition, DRD2TaqIa and 5-HTTLPR, associated with striatal D(2) receptor density and serotonin uptake, respectively, were assessed. Participants received placebo or a selective DA-D(2) receptor blocker (sulpiride, 200 mg) and performed a Flanker task. EEG was recorded and decomposed into independent brain components (ICs) using independent component analysis. After errors, participants displayed (1) a negative deflection in ICs source-localized to the proximity of the anterior midcingulate cortex [IC-error-related negativity (IC-ERN)], (2) increased midcingulate cortex IC power in the delta/theta frequency range, and (3) slowing in the subsequent trial [posterror slowing (PES)]. Importantly, all, IC-ERN, delta/theta power, and PES were modulated by COMT × Substance interactions such that the Val allele predicted elevated IC-ERN, delta/theta power, and PES after placebo; this association was reversed under sulpiride. Because low doses of sulpiride presumably increase PFC DA levels, the COMT × Substance interaction supports the hypothesis that low (Val, placebo) and high (Met, sulpiride) versus medium (Val, sulpiride; Met, placebo) DA levels elevate reactivity to errors. Consistent with an influence of serotonin on PFC DA, the COMT × Substance interaction was modulated by 5-HTTLPR.

Figure 1.

Postulated relationship between prefrontal dopamine level, COMT, and relative D1- versus D2-receptor activation as previously described by Durstewitz and Seamans (2008). Due to enhanced relative D2 receptor activation in Val+ versus Met/Met carriers, we hypothesized increased error reactivity in Val+ versus Met/Met carriers. By increasing PFC dopamine activity through presynaptic D2 receptor blockade, sulpiride (200 mg) is predicted to shift Val+ carriers into medium and Met/Met carriers into high dopamine levels (red arrows), resulting in reduction or enhancement of error reactivity, respectively.

Figure 2.

Cluster-mean IC scalp topographies for outlier (Cluster 1, top left) and other IC clusters. Cluster 9 (red box) was used for analyses of error-related brain activity in the present study.

Figure 3.

Interactions of Substance and COMT on neural and behavioral error-processing correlates. a, Grand average response-locked event-related potentials for a medial frontal IC-cluster following erroneous button presses for Val+ (gray lines) and Met/Met (black lines) carriers who had received placebo (thick lines) or sulpiride (thin lines). A latency of 0 ms corresponds to the time of the button press. Independent component ERPs were normalized, dividing by the root mean square over the component scalp map projection to all channels before averaging (see Materials and Methods). A standard template brain image (Montreal Neurological Institute) indicates the region of maximum concentration measured as equivalent dipole density for this IC cluster. b, Event-related spectral perturbation plot showing cluster mean differences in log spectral power relative to log spectral power at button press, separately for the two (Met/Met vs Val+) by two (placebo vs sulpiride) matrix of subgroups. A latency of 0 ms corresponds to the time of the button press. For each subgroup, the mean estimated dipole position is also plotted in the upper right corner of the event-related spectral perturbation plot. c–e, Bar plots indicating means (and SEMs) of reaction-time slowing in the subsequent trial following errors (c), peak IC-cluster error-related negativity (d), and average power increase from 1.5 to 8 Hz (e) in the placebo (white) or sulpiride (gray) groups.

Figure 4.

Interactions of Substance, COMT, and 5-HTTLPR on peak IC cluster event-related potentials. Bar plots indicating means (and SEMs) of peak IC cluster event-related potentials following errors in the placebo (white) or sulpiride (gray) groups.