Functional Organization of Prefrontal Cortex and its role in the neural systems of higher cognition.
Research in both humans and nonhuman primates has demonstrated that the prefrontal cortex has a complex functional anatomical organization. Attempts to elucidate the precise nature of this organization, however, have met with only limited success. There is evidence for organizational schemes based both on the type of information being processed and on the processing and cognitive control demands of the current task. Research suggests that there may even be an interaction between information type and processing demands. Using a variety of carefully controlled behavioral paradigms combined with functional magnetic resonance imaging, we have demonstrated that the human prefrontal cortex, like that of the monkey (e.g. Levy and Goldman-Rakic, 2000), is organized according to the type of information being maintained (Courtney et al., 1996, 1998; Sala et al, 2003; Rämä et al, 2004 and in press). However, it now appears that this dorsal/ventral functional topography may depend on a general spatial/nonspatial dissociation, rather than the traditional “where/what” distinction, with objects having a distributed representation across both dorsal and ventral frontal regions (Sala et al, 2003 and conference abstracts Yee et al., 2003 and Roe et al., 2004). In addition, we have investigated the representation of the link between an object and its location with both behavioral and imaging experiments. We have developed a computational model that reconciles evidence for domain specificity with evidence for single cells with integrated object-in-location selectivity, explaining the neuroimaging results in terms of the activities and interactions of individual cells (Sala 2003, Ph.D. Dissertation, abstracts, manuscript under review). Within a single information domain, our research has shown that the prefrontal cortex is organized hierarchically with posterior areas responding to both the stimulus presentation and maintaining that activity during a working memory delay, and anterior areas responding only during the delay period after the stimulus is removed from view (Courtney et al., 1997). We are further investigating the nature of this hierarchy with the hypothesis that the anterior-posterior axis of prefrontal organization is based on the contextual and control requirements of the task. One ongoing set of experiments directed toward this goal contrasts maintaining old information versus controlling (updating) the contents of working memory with newly relevant information, either from perceptual input (Roth et al., 2003 abstract) or retrieval from long-term memory storage. These control mechanisms may not be specific to working memory as some similar cortical regions appear to be activated by shifts of both spatial and object-based attention (Yantis et al., 2002; Serences et al., 2004)
Functional Specialization, Distributed Representation, and Interactions among Brain Areas.
Cortical regions do not act in isolation. The organization of the cerebral cortex, including the prefrontal cortex, can be described neither in terms of strict functional modularity nor in terms of equipotentiality. The kinds of computations that are best performed by a cortical region are determined by the cellular, molecular, and connectivity properties of that region. Those computations, however, can subserve multiple functions in the service of performing a cognitive task. A region’s functional selectivity and it’s role in the particular task at hand can be influenced by interactions with other brain regions. One advantage of a brain imaging technique such as fMRI is the ability to simultaneously record activity patterns across the entire brain. This allows one to examine issues of functional connectivity among the large number of cortical and subcortical regions involved in such cognitive tasks. Another way to investigate how brain areas normally interact is to observe what happens when that communication is disrupted. We are currently studying patients with Multiple Sclerosis (MS), an immune-mediated demyelinating disease. Preliminary data suggests that even in patients with no measurable cognitive deficits, cortical regions normally involved in working memory tasks can activate normally under some conditions but show abnormal activation (either increased or decreased) under other conditions, depending on the task and the locations of the white matter lesions (Sala et al, 2003; Morgen et al., 2003; Courtney and Sayala, 2004 conference abstracts).
Plasticity and Development
Our experience with these MS patients who can have normal cognitive performance despite dramatically altered cortical activation patterns has suggested a remarkable amount of adaptability or plasticity in the adult brain for performance of cognitive tasks in the face of disrupted communication between cortical regions. We are currently pursuing studies looking at both short-term (within 1 hour) and long-term (months and years) changes in activation patterns in both normal adults and in patients with MS (Sayala et al., 2003 abstract, manuscript in preparation). By following people with a relapsing-remitting course of MS longitudinally, we will hopefully be able to observe the timecourse, extent and limitations of this type of plasticity in cognitive neural systems. In addition to this interest in adult plasticity, I have ongoing collaborations that are aimed at elucidating (1) the development of mechanisms of cognitive control in the prefrontal cortex, and (2) the development and nature of the dorsal/ventral spatial/nonspatial functional anatomical dissociation. We have pursued these questions by studying children with Attention Deficit Hyperactivity Disorder (Mostofsky et al, 2003) and people with Williams Syndrome (O’Hearn Donny et al, 2004, abstract), respectively.