The Cross-Function Approach to Imaging Cognition

During the last decade, the field of functional neuroimaging of cognition has grown exponentially. From a handful of studies in the early 1990s, this research domain expanded to more than 800 studies by the early 2000s. Today, positron emission tomography (PET) and functional MRI (fMRI) studies cover almost every aspect of human cognition, from motion perception to moral reasoning. If each study is seen as a tree, the field has grown from minimal vegetation to a luxuriant tropical forest in less than ten years. Yet, functional neuroimaging researchers sometimes focus exclusively on their own cognitive domain and do not see the forest through the trees. The goal of the present chapter is to call attention to the forest, that is, to what many functional neuroimaging studies of cognition have in common. When we say that most researchers are focused on the trees, we refer to the fact that the vast majority of functional neuroimaging studies investigate a single cognitive function, such as attention, working memory, or episodic memory. Yet, with the accumulation of functional neuroimaging data, it has become obvious that the brain is not organized like a cognitive psychology textbook with dedicated systems for perception, attention, working memory, episodic memory, and so forth. Instead, the neural correlates of cognitive functions overlap considerably, with each brain region being involved in a variety of cognitive functions. What cognitive processes do these common regions mediate? By comparing patterns of brain activity across different cognitive functions, answers to this question can be generated. -------Figure 1 about here -------The matrix in Figure 1 illustrates the difference between the traditional within-function approach and the cross-function approach we are advocating in this chapter. Let us assume that in functional neuroimaging studies Cognitive Function A typically is associated with activations in Brain Regions 1 and 3, Cognitive Function B with activations in Brain Regions 2 and 3, and Cognitive Function C with activations in Brain Regions 1 and 2. In the standard within-function approach, functional neuroimaging researchers are primarily concerned with one cognitive function and interpret activations in relation to this particular function. Thus, in a situation like the one depicted in Figure 1, researchers of Function A would attribute the activation of Region 1 to a certain aspect of Function A, whereas researchers of Function C would attribute the activation of the same region to a certain aspect of Function C. For instance, left ventrolateral prefrontal cortex activations have been attributed to language processes by language researchers, to working memory processes by working memory researchers, to semantic memory processes by semantic memory researchers, and so forth (Cabeza & Nyberg, 2000). In contrast with the within-function approach, the crossfunction approach focuses on the columns of the matrix rather than on the rows, and asks questions about the functional role of a brain region (e.g., Region 1) that is recruited by different cognitive functions (e.g., Functions A and C). Thus, the basic question the cross-function approach asks is why the same brain region is recruited by different cognitive functions. There are at least three possible answers to this question. First, according to a sharing view, the common region is involved in cognitive operations that are recruited by different cognitive functions. In the case of Figure 1, the sharing view would argue that Region 1 mediates processes that are engaged both by Function A and C. A reductionistic interpretation of the sharing view