Purchase this article with an account.
Kunjan D. Rana, Amanda C. Del Giacco, Elizabeth A. Kruse, Valentinos Zachariou, Leslie G. Ungerleider; Time-resolved functional networks involved in discriminating faces with configural differences. Journal of Vision 2020;20(11):743. doi: https://doi.org/10.1167/jov.20.11.743.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Human face recognition is a critical cognitive operation that is performed within a fraction of a second. Facial recognition is thought to be mediated by a network of face-selective areas along the ventral visual stream. Zachariou and colleagues implicated the dorsal visual stream in processing configural facial features, namely, the spatial arrangement of face components (Zachariou et al, 2016). Here, we measured the timing and interaction between the dorsal and ventral visual streams during a face discrimination task. Subjects viewed a sequence of faces during a magnetoencephalography (MEG) scan session, organized into 20 blocks of 21 faces. Faces had either configural, featural, or no difference. Subjects pressed a button if the presented face differed from the previous face. Two face-selective regions (occipital face area (OFA) and fusiform face area (FFA)) and one dorsal stream region (intraparietal sulcus (IPS)) were identified using functional localizers. Dynamic Granger Causality (DGC) revealed bidirectional connectivity between OFA and IPS (100-240 ms) and between FFA and IPS (220-310 ms) during configural face differences but not featural differences, whereas DGC revealed a unidirectional connection from OFA to FFA during both configural and featural differences (100-240 ms). Ding and colleagues have proposed that bidirectional Granger connectivity can be caused by a number of interactions, such as being driven by a third source (Ding et al, 2006). We conducted a spotlight search using conditional Granger Causality to locate regions that modulated bidirectional connections. We found a network between 100-240 ms involving early visual cortex (EVC), OFA, and IPS, and a second network between 220-310 ms involving the anterior inferior temporal cortex (aIT), dorsolateral prefrontal cortex (DLPFC), IPS, and FFA. We suggest that the first network processes facial features, whereas the second network performs face discrimination, due to involvement of anterior face areas (FFA and aIT) together with prefrontal cortex.
This PDF is available to Subscribers Only