Although task switching is often considered one of the fundamental abilities underlying executive functioning and general intelligence, there is little evidence that switching is a unitary construct and little evidence regarding the relationship between brain activity and switching performance. We examined individual differences in multiple types of attention shifting in order to determine whether behavioral performance and fMRI activity are correlated across different types of shifting. The participants (nϭ 39) switched between objects and attributes both when stimuli were perceptually available (external) and when stimuli were stored in memory (internal). We found that there were more switchrelated activations in many regions associated with executive control-including the dorsolateral and medial prefrontal and parietal cortices-when behavioral switch costs were higher (poor performance). Conversely, activation in the ventromedial prefrontal cortex (VMPFC) and the rostral anterior cingulate was consistently correlated with good performance, suggesting a general role for these areas in efficient attention shifting. We discuss these findings in terms of a model of cognitive-emotional interaction in attention shifting, in which reward-related signals in the VMPFC guide efficient selection of tasks in the lateral prefrontal and parietal cortices.
WAGER, JONIDES, SMITH, AND NICHOLStasks can be dissociated on the basis of three factors: (1) Among what representations is attention shifted? (2) In what locus are the items between which attention is shifted stored-in WM or in perceptual buffers? (3) And finally, when does the shift occur-when an actual stimulus is presented and a behavioral response is required, or in advance of stimulus presentation and response selection?
What Representations and What Locus of Storage?Theories of switching have combined evidence on switch costs from studies varying the cue type (Coull, Frith, Buchel, & Nobre, 2000), the target tasks (Allport, Styles, & Hsieh, 1994;Rubinstein, Meyer, & Evans, 2001), and whether switching occurs among the tasks themselves (i.e., both stimulus and response sets), operations (Spector & Biederman, 1976), stimulus-response mapping rules (Rubinstein et al., 2001), objects stored in WM (Garavan, 1998), object attributes (Owen, Roberts, Polkey, Sahakian, & Robbins, 1991;Rogers & Monsell, 1995), or combinations of these. The question of consistency across types is critical if particular shifting tasks are used as measures of an executive ability.In this study, we combined two types of evidence that have been considered for understanding the relationships between types of shifting: individual differences in performance and brain-imaging activation. In individualdifferences studies, the basic measures of performance are the reaction time (RT) and error rate costs for trials in which a shift of attention is required, as compared with trials with no shift (i.e., shift-no shift; see, e.g., Garavan, 1998;Gopher, Armony, & Greenshpan, 2000;Meiran, Chorev, & Sapir, 2...