Neural mechanisms of interference control underlie the relationship between fluid intelligence and working memory span

Abstract

Fluid intelligence (gF) and working memory (WM) span predict success in demanding cognitive situations. Recent studies show that much of the variance in gF and WM span is shared, suggesting common neural mechanisms. This study provides a direct investigation of the degree to which shared variance in gF and WM span can be explained by neural mechanisms of interference control. The authors measured performance and functional magnetic resonance imaging activity in 102 participants during the n-back WM task, focusing on the selective activation effects associated with high-interference lure trials. Brain activity on these trials was correlated with gF, WM span, and task performance in core brain regions linked to WM and executive control, including bilateral dorsolateral prefrontal cortex (middle frontal gyrus; BA9) and parietal cortex (inferior parietal cortex; BA 40/7). Interference-related performance and interference-related activity accounted for a significant proportion of the shared variance in gF and WM span. Path analyses indicate that interference control activity may affect gF through a common set of processes that also influence WM span. These results suggest that individual differences in interference-control mechanisms are important for understanding the relationship between gF and WM span.

Figure 1

Venn diagrams illustrating variance-partitioning method used to divide variance in gF, WM span, and interference control measures into common and shared portions. Figure 1a illustrates the general method conceptually. Figure 1b shows the variance portions related to lure accuracy. Figure 1c shows the variance portions related to lure activity factor scores. N.B., size of the overlapping circles is not drawn in proportion to the amount of variance explained by those portions.

Figure 2

Brain regions in which lure-related activity correlates with gF, WM span, and lure accuracy. Color of regions reflects z-statistic for correlation between lure-related activity and WM span. Scatterplots demonstrate the correlations of the lure activity factor scores with gF, WM span, and lure accuracy respectively.

Figure 3

Path model predicting gF with indirect paths from lure activity and lure accuracy through WM span