Relationships Among Fluid Intelligence, Working Memory Span, and Brain Activity During High -Interference Trials

Gregory C. Burgess, Jeremy R. Gray, Andrew R. A. Conway, Todd S. Braver

Abstract Text : Fluid intelligence (gF) is thought to be a major dimension of individual variation in cognitive ability, reflecting the capacity to reason and solve problems in novel domains. Despite its importance, currently little is known about the specific neural mechanisms that cause individual differences in gF. Two recent studies from our laboratory have related gF to brain activity during high -interference trials, in which the maintenance of task-relevant information was critical for accurate performance. Interestingly, the experimental literature has suggested that variation in gF can be explained by variation in working memory (WM) span, and more specifically, by the ability to actively maintain goal - relevant information in WM in the face of interference. The current study extended our previous work by examining the relationship between gF and WM span, and the degree to which that relationship is explained by neural activity under high -interference conditions. 120 participants performed the n -back WM task (3 -back condition) while being scanned with 3T event-related fMRI. Non-target trials were classified as either low or high interference (recent repeats). Offline, participants completed two standard measures of gF and four measures of WM span. The two latent constructs derived from these measures were highly correlated. Neuroimaging analyses examined the statistical relationship of neural activity on high -interference trials to the gF and WM constructs. In particular, we examined whether high -interference activity in specific brain regions (e.g., bilateral PFC) statistically mediated the gF ­ WM span relationship. The results have important implications for understanding the neural and cognitive mechanisms of gF and WM span.