Spatiotemporal Imaging of Semantic Memory

描述（由申请人提供）：该应用侧重于神经系统

DESCRIPTION (provided by applicant): The application focuses on the neural substrates of semantic memory, which is a human's knowledge of the world. During the last two decades, a number of articles have described category-specific deficits in patients with damaged/diseased brains. These deficits involve selective degradation of one or more categories of knowledge. For instance, some patients exhibit loss of knowledge of animals, but not tools, while others show the reverse. During the 1990s, prominent neuroimaging studies using PET and fMRI have shown category-specific patterns of brain activation in healthy subjects. Together, these two sets of findings have spurred debate (but no definitive answers) regarding the nature and origin of category-specific effects. This application tackles this question by supplementing extant neuropsychological and neuroimaging research with data from event-related brain potentials (ERPs) recorded with a dense array of 128 electrodes analyzed with new methods for localizing the sources of scalp-recorded electrical activity in the brain co-registered with MRI. The basic idea is that in order to understand category-specific effects in both patient and nonpatient populations, it is important to determine what parts of the brain are actually representing semantic information. Static images of the parts of the brain that are involved in processing different categories of semantic knowledge are insufficient for determining the critical storage sites, because such neuropsychological and neuroimaging evidence does not easily distinguish areas that represent the information from areas that later perform other operations on this information. To localize semantic representation, it is necessary to isolate the access mechanisms that yield this information from the representation sites. This can be done using ERPs, because the superb temporal resolution of this technique allows one to focus on neural activity during the early time-window of processing during which semantic access predominates. Using new technologies, these mechanisms can be analyzed and localized in the brain. The proposed research involves applying this approach to a number of semantic memory paradigms in order to provide spatiotemporal information to help interpret neuroimaging/neuropsychological studies. Furthermore, the project will use such results to inform the development of connectionist attractor models of semantic memory. Isolating, localizing, and theoretically characterizing semantic representation and access should help in the understanding and diagnosis of disorders of semantic memory and should suggest possible therapies.