Abstract
Numerous neuroimaging studies have identified an extended network of face responsive regions in the human brain. For example, robust activations for face stimuli have been found along the fusiform gyrus, the superior temporal sulcus, and in the anterior temporal lobe. Recent functional MRI (fMRI) studies in humans and monkeys have also reported face patches in the lateral prefrontal cortex in the absence of working memory demands, confirming findings from early electrophysiological studies in monkeys. Further investigation has demonstrated that this face activation is confined to the right inferior frontal junction (IFJ), with stronger response elicited by pairs of eyes than faces with eyes covered (Chan & Downing 2011; Chan 2013). However, many of these studies have only recruited a handful of participants. Here, we aimed to further investigate the topographical location of this prefrontal face activation and its response properties with a big dataset. Using fMRI data of ~500 healthy adults from the Human Connectome Project, our preliminary results showed activation beyond the IFJ when contrasting faces with objects. Two additional areas also elicited strong activation for faces, one above the IFJ near to the frontal eye field, another anterior to the IFJ. These strong activations were observed across both hemispheres. Individual subject region-of-interest analysis showed that these prefrontal areas produced different patterns of selectivity, suggesting that there are pockets of strong responses to visual categories, which may reflect some sort of underlying organizational principle for visual processing. Resting-state functional connectivity analysis with IFJ as a seed region showed a much stronger correlation with the superior temporal sulcus and middle temporal gyrus, indicating that these lateral regions may extract similar visual properties during face processing. Overall, this study verifies the presence of visual category information in the human prefrontal cortex in a larger population.
Meeting abstract presented at VSS 2016