"How Does the Brain Recognize Faces?" by Lindsey Jay
Facial recognition is one of the most important skills for successful daily interactions, and it is an automatic process that we often take for granted. However, the mechanisms underlying facial recognition are quite complex, and scientists have been researching how such an intuitive task works. One of the simplest questions to tackle was whether humans had a specific region in the brain responsible for identifying faces or whether multiple areas of the brain categorized different aspects of facial features, such as eye curvature or nose shape. This debate has respectively been separated into the domain-specific and domain-general hypotheses.
Research using functional magnetic resonance imaging (fMRI) has supported the domain-specific hypothesis by identifying a region of the human extrastriate cortex, the fusiform face area (FFA) in the fusiform gyrus (FG), that is significantly more active when human subjects viewed faces than when they viewed other common objects. Kanwisher et al. at Harvard University first tried identifying which brain regions were significantly more active when viewing faces versus other objects. They found that, in 12 out of 15 subjects, an area in the right FG was highly activated. On the other hand, a completely separate brain region, the parahippocampal gyrus, was active when subjects viewed non-facial objects. To further investigate the role of the FG, they tested the response to intact and completely rearranged two-tone faces to investigate whether the distinct facial arrangement rather than mixed elements of a face elicited the FG response. They also tested the response to faces versus houses to see if the neural activity was purely for facial recognition and not just objects. In both cases, they found that the FG reliably responded more strongly to the intact face stimuli. This confirmed that FG activation was due the identification of faces rather than general objects.
Yet another study by Zhang et al. at Beijing Normal University supports the domain-specific hypothesis with the finding that the FFA is engaged in a holistic rather than parts-based representation of faces. Participants were presented with either anatomically correct pictures of faces or pictures of faces with the correct facial features but in the wrong anatomical position (“scrambled”). Upon analysis, they found that participants were less accurate in discriminating facial parts when the face was scrambled than with an anatomically correct face. This result suggested that both the complete presence of facial features and their overall holistic arrangement was crucial for activation of the FFA.
While these studies merely indicate that the FG is correlated with facial recognition, another study by Parvisi et al. at Stanford University provided concrete evidence that the FFA plays a causal role in facial recognition. They specifically studied the brain activity of a single patient implanted with subdural electrodes in the right inferior temporal lobe. Two FG face-selective regions (mFus-faces and pFus-faces) were identified with fMRI after the subject was shown pictures of faces, which anatomically and functionally correlated with electrocorticography (ECoG) data. When they delivered electrical charge via electrical brain stimulation (EBS) to the pair of implanted electrodes, the patient reported profound face-specific distortion when viewing real faces. Electrical charge was delivered to the electrodes in sham (0 mA) and real trials (2-4 mA). The electrodes were stimulated when the subject viewed either the face of specific individuals such as a doctor or objects in the room such as a television. The subject reported vivid distortions of faces during real electrode stimulation but not during the sham procedures. He further reported the face was a distortion rather than a morphing into an intact face of another person. The subject did not report such distortions when the electrodes were stimulated while he viewed other, non-facial objects in the room. While this provides compelling evidence towards the domain-specific hypothesis, this was a case-study experiment and the results may not be true of all human subjects. Therefore, a next step would be to perform this study in more subjects to strengthen their findings.
We have made significant progress untangling this simple question of whether facial recognition is domain-specific or domain-general, with overwhelming research supporting the former, paving the road for further findings in facial recognition mechanisms. While scientists have already discovered a great deal about how we are able to recognize faces, it is still unclear to what extent the FG plays a role in facial recognition (does it help us recognize face stimuli in general, or does it also store familiar faces?). Thus, much about the underlying facial recognition mechanisms are still waiting to be discovered.
 Kanwisher, Nancy, and Galit Yovel. "The fusiform face area: a cortical region specialized for the perception of faces." Philosophical Transactions of the Royal Society B: Biological Sciences. December 29, 2006. Accessed April 15, 2017. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1857737/.
 Kanwisher, Nancy, Josh McDermott, and Marvin M. Chun. "The Fusiform Face Area: A Module in Human Extrastriate Cortex Specialized for Face Perception." Journal of Neuroscience. June 01, 1997. Accessed April 15, 2017. http://www.jneurosci.org/content/17/11/4302.long#sec-6.
 Zhang, Jiedong, Xiaobai Li, Yiying Song, and Jia Liu. "The Fusiform Face Area Is Engaged in Holistic, Not Parts-Based, Representation of Faces." PLOS ONE. July 6, 2012. Accessed April 15, 2017. http://journals.plos.org/plosone/article?id=10.1371%2Fjournal.pone.0040390.
 Parvizi, Josef, Corentin Jacques, Brett L. Foster, Nathan Withoft, Vinitha Rangarajan, Kevin S. Weiner, and Kalanit Grill-Spector. "Electrical stimulation of human fusiform face-selective regions distorts face perception." The Journal of neuroscience: the official journal of the Society for Neuroscience. October 24, 2012. Accessed April 15, 2017. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3517886/.