| The process of recognizing and identifying a familiar person (face or name) represents a basic and important activity of daily social interaction. Both clinical lesion studies and recent functional neuroimaging studies have contributed to the understanding of the underlying psychological processes and brain regions involved in person recognition and identification. These studies have consistently found a widespread bilateral neural network that includes the posterior cingulate, anterior and medial temporal lobes, hippocampal and parahippocampal regions, and prefrontal regions (Damasio et al., 1996; Fink et al., 1996; Craige & Hanley, 1997; Gorno-Tempini et al., 1998; Leveroni et al., 2000; Maddock et al., 2001; Seidenberg, et al., 2002; & Bernard et al., 2004). We have come to refer to these cognitive operations and neural underpinnings as the person identity network (PIN). Bruce and Young (1986) proposed a model of face recognition and identification, which has provided a valuable conceptual framework for the broader investigation of the PIN.; The goal of this study was to investigate the influence of pre-existing semantic knowledge on the neural system mediating person identification. It was hypothesized that the fMRI Blood Oxygenation Level Dependent (BOLD) signal would co-vary with semantic knowledge. Specifically, we expected a higher signal response for high semantic knowledge than low semantic knowledge, particularly in the anterior temporal lobes, frontal cortex, posterior cingulate, and hippocampal region.; Thirty-two healthy, elderly individuals were asked to discriminate familiar from unfamiliar names in an event-related fMRI design. All participants underwent a neuropsychological assessment, APOE genotype testing, a fMRI scanning procedure, and a post-scanning questionnaire. The BOLD signal changes associated with extent of semantic knowledge ratings were extracted for analysis.; A voxel-wise analysis was performed to determine the fMRI network activated by the name recognition task and amount of semantic knowledge for each famous name. Voxel-wise t-tests for High semantic knowledge -- Low semantic knowledge produced two significant clusters. The clusters indicated two regions that were left lateralized: the posterior cingulate (962 ml) and a second region that included the anterior cingulate and medial frontal gyrus (692 ml).; A functional region of interest (ROI) analysis was also conducted as a follow-up to the voxel-wise analysis to more directly evaluate possible differences across high and low semantic knowledge in MR signal intensity, as reflected by the area under the curve (AUC) of the hemodynamic response function. A conjunctive map of the high and low semantic knowledge conditions identified six functional ROIs (p<.003, volume > 0.200 ml) including the left inferior frontal gyrus, angular gyrus, middle frontal gyrus and the fusiform gyrus, while in the right hemisphere the medial frontal gyrus and posterior cingulate were activated. Paired t-tests of high minus low semantic knowledge revealed no significant response differences in the area under the curve across the six regions.; Contrary to our prediction, results indicated that fMRI BOLD activity did not differ as a function of semantic knowledge. Possible reasons for lack of positive findings include low levels of overall semantic knowledge, and the temporal incongruity of assessing semantic knowledge and BOLD signal activation. However, consistent with previous literature, we found that famous names produced more activation than unfamiliar famous names. The famous names minus non-famous names contrast produced activation in a bilateral network consisting of the posterior cingulate, anterior and medial temporal lobes, and prefrontal regions (Gorno-Tempini et al., 1998; Douville et al., 2005; Nielson et al., 2006; Woodard et al., 2007). In summary, our current findings suggest that amount of semantic knowledge, at least as measured here, does not pl... |
