Comparison of brain imaging studies reveals that many cognitive functions tend to recruit overlapping neural regions (for reviews, see Cabeza & Nyberg Reference Cabeza and Nyberg2000; Duncan & Owen Reference Duncan and Owen2000; Naghavi & Nyberg Reference Naghavi and Nyberg2005). It is possible that the apparent commonalities reported in such between-studies assessments, at least in part, reflect activation of adjacent but distinct regions. However, several within-study PET and fMRI reports have also demonstrated overlap in activation patterns for different cognitive functions, such as attention, working memory, and episodic memory retrieval (e.g., Braver et al. Reference Braver, Barch, Kelley, Buckner, Cohen, Miezin, Snyder, Ollinger, Akbudak, Conturo and Petersen2001; Cabeza et al. Reference Cabeza, Dolcos, Graham and Nyberg2002; Reference Cabeza, Dolcos, Prince, Rice, Weissman and Nyberg2003; LaBar et al. Reference LaBar, Gitelman, Parrish and Mesulam1999; Nyberg et al. Reference Nyberg, Forkstam, Petersson, Cabeza and Ingvar2002; Ranganath et al. Reference Ranganath, Johnson and D'Esposito2003). Common activations are particularly prominent in interconnected cortical regions in the frontal and parietal lobes, including the dorsolateral prefrontal cortex and posterior parietal cortex (see our Fig. 1). Reviewing 37 neuroimaging studies, J&H conclude that a very similar large-scale network mediates performance in tasks involving intelligent thought and reasoning. This conclusion suggests that an even broader range of conditions activate the fronto-parietal circuit, although within-study comparisons are needed to exactly and reliably define these similarities. The fronto-parietal recruitment is particularly salient for tasks having a high g-loading, thereby further strengthening the association between fronto-parietal activity and general-purpose higher-order cognitive functions.
Peaks of activation in fronto-parietal areas associated with: (A) attention, (B) working memory, (C) episodic retrieval, and (D) conscious perception, extracted from 47 fMRI and PET studies; blank areas indicate the regions with highest level of common regional brain activity across functions, namely bilateral BA 7 and BA 40, left BA 6, and right BA 9 (after Naghavi & Nyberg Reference Naghavi and Nyberg2005; reprinted with permission from Elsevier).
What does the convergence of diverse cognitive functions in a large-scale fronto-parietal network imply? From a general point of view, common fronto-parietal activations reveal an important feature of how cognitive processes are distributed in the brain. Computational systems are traditionally thought of as devices in which distinct components are devoted to specific functions, with minimal if any overlapping localization of functions. The brain, however, appears to be designed in such a way that particular regions are concurrently involved in numerous cognitive functions (Duncan Reference Duncan2001; McIntosh Reference McIntosh2000). In fact, compelling evidence from neuroimaging and neuropsychological studies suggests that the frontal lobe, particularly the dorsolateral prefrontal cortex, and the parietal cortex, especially the intraparietal sulcus, collaborate tightly for directing a wide range of higher cognitive functions as well as sensory-motor processes (Bush et al. Reference Bush, Vogt, Holmes, Dale, Greve, Jenike and Rosen2002; Collette & Van der Linden 2002; Culham & Kanwisher Reference Culham and Kanwisher2001). The notion of multifunctionality of the fronto-parietal network is also supported by single-cell studies in primates, which demonstrate the co-presence of neurons with different processing operations (Funahashi et al. Reference Funahashi, Bruce and Goldman-Rakic1989; Reference Funahashi, Bruce and Goldman-Rakic1990), and/or the presence of neurons with multiple processing operations (Fuster et al. Reference Fuster, Bauer and Jervey1982; Quintana & Fuster Reference Quintana and Fuster1999; Rosenkilde et al. Reference Rosenkilde, Bauer and Fuster1981) in regions in the prefrontal cortex and the parietal cortex. For example, in monkeys it has been shown that neurons in the lateral intraparietal area are activated across a wide variety of conditions, including visual, attentional, memory, and saccade-related tasks (Colby et al. Reference Colby, Duhamel and Goldberg1996).
In spite of the apparent diversity of cognitive functions that correlate with fronto-parietal activity, these functions may fit into a unifying conceptual framework for integration and control of information. Such a capability is critical for optimal recruitment of internal resources to exhibit goal-directed behavior relevant to ever-changing environmental requirements (Miller & Cohen Reference Miller and Cohen2001). Examples of integration and control processes are: multimodal convergence of behaviorally relevant information in coherent representations; selective enhancement or inhibition of specific representations through feedback mechanisms; maintenance of information in a buffer system via sustained activity; and manipulation of information according to the cognitive demands. All of these processes should be carried out by a central system that has extensive access to both sensory and motor representations, and the fronto-parietal network is at an ideal site in the brain to subserve this end. Nodes of the fronto-parietal network are thoroughly and reciprocally connected with each other, as well as with other association cortices and subcortical areas, a property that allows widespread access to perceptual and motor representations at different levels. With this unique connectivity pattern, on the one hand, and specialization in a wide variety of higher-order processing operations, on the other hand, the fronto-parietal network can function as the source of integration and top-down control in the brain, orchestrating perception, thought, and action in accordance with internal goals.
Hence, J&H's conclusion that intellectual demands activate the fronto-parietal network is consistent with a general role of the fronto-parietal network in integration and control processes. However, it should be noted that a number of relatively basic functions, such as conscious perception (Rees et al. Reference Rees, Kreiman and Koch2002), top-down attention (Pessoa et al. Reference Pessoa, Kastner and Ungerleider2003), and eye movement control (Muri Reference Muri2005), are also associated with activity in the fronto-parietal network. Therefore, whereas J&H argue for viewing the parieto-frontal network as a neural signature of intelligence, the association between fronto-parietal activity and intelligence may reflect only a specific, though important, aspect of a much more general role of the fronto-parietal network. As an integrative model, the fronto-parietal network can be seen as a core system equipped with diverse mechanisms that allow integration and control of distributed patterns of neural activity throughout the brain. Given the distribution of the fronto-parietal network in posterior and anterior parts of the brain, as well as its connectivities and functional specializations, this network might be conceived of as a backbone, located at the top of the hierarchical organization of the brain, by which the otherwise fragmented pieces of information as well as sensory-motor and cognitive processes are integrated and managed. Thereby, integrative action in the fronto-parietal cortex provides a cure for a scattered mind.
