COMT val158met genotype affects recruitment of neural mechanisms supporting fluid intelligence

Abstract

Fluid intelligence (g(f)) influences performance across many cognitive domains. It is affected by both genetic and environmental factors. Tasks tapping g(f) activate a network of brain regions including the lateral prefrontal cortex (LPFC), the presupplementary motor area/anterior cingulate cortex (pre-SMA/ACC), and the intraparietal sulcus (IPS). In line with the "intermediate phenotype" approach, we assessed effects of a polymorphism (val(158)met) in the catechol-O-methyltransferase (COMT) gene on activity within this network and on actual task performance during spatial and verbal g(f) tasks. COMT regulates catecholaminergic signaling in prefrontal cortex. The val(158) allele is associated with higher COMT activity than the met(158) allele. Twenty-two volunteers genotyped for the COMT val(158)met polymorphism completed high and low g(f) versions of spatial and verbal problem-solving tasks. Our results showed a positive effect of COMT val allele load upon the blood oxygen level-dependent response in LPFC, pre-SMA/ACC, and IPS during high g(f) versus low g(f) task performance in both spatial and verbal domains. These results indicate an influence of the COMT val(158)met polymorphism upon the neural circuitry supporting g(f). The behavioral effects of val allele load differed inside and outside the scanner, consistent with contextual modulation of the relation between COMT val(158)met genotype and g(f) task performance.

Figure 1.

Example test items for each task. Each item consisted of 4 display elements (drawings or letter sets), and the task was to identify the element that in some sense mismatched or differed from the others. Materials for the high g_f tasks were adapted with permission from a standard nonverbal test of g_f, Cattell Culture Fair, Scale 2 Form A and Scale 3 (Institute for Personality and Ability Testing 1973) and a standard letter-based problem-solving task, Letter Sets from the ETS kit of factor-referenced tests (Ekstrom et al. 1976). The high g_f loading of these tasks was previously established (Wothke et al. 1991; Duncan et al. 2000). The low g_f items were structurally similar but with a minimal problem-solving component. Participants were asked to select the only nonidentical item for the low g_f spatial task and to select the string not in alphabetic order for the low g_f verbal task.

Figure 2.

Neural activation associated with high g_f versus low g_f task performance. (A) Significant activations at a whole-brain fdr threshold of P < 0.05, rendered onto the canonical T₁-weighted brain image of SPM99. (i) Spatial high g_f − spatial low g_f. (ii) Verbal high g_f − verbal low g_f. (B) Activation peaks. For significant (fdr, P < 0.05) clusters of any size that overlap the a priori specified ROIs, the table gives peak voxel within the ROI. For other brain areas, peaks are reported only for clusters of 20 or more significant voxels. OFC: orbitofrontal cortex.

Figure 3.

Correlation between number of COMT val alleles possessed (0: met/met, 1: val/met, and 2: val/val) and activation for high g_f versus low g_f conditions for the spatial problem-solving task (A: right DLPFC, B: dorsal ACC, C: pre-SMA, and D: left IPS) and the verbal problem-solving task (E: right DLPFC, F: right FO/AI, G: pre-SMA, and H: left IPS). Left: Activations thresholded at P < 0.05 svc are overlaid on the canonical T₁ SPM99 brain. Right: Individual BOLD responses were extracted from the voxel with the highest Z value inside the ROI and plotted against number of COMT val alleles possessed.