Functional Anatomy of Face Processing in the Primate Brain

National Institute Of Mental Health

Investigators

Linked publications & trials

Abstract

Lesions of inferior temporal (IT) cortex in humans can result in the syndrome termed prosopagnosia, an inability to recognize familiar faces. Single-cell recordings from IT cortex of monkeys have revealed the existence of neurons that are selectively activated by visual images of faces. Additionally, fMRI of both human and monkey brains has demonstrated face-selective regions, in which the fMRI signal evoked by faces is greater compared to that evoked by non-face objects. Together, these studies point to specialized neural machinery in the primate brain for processing faces. Thus far, our group and others have shown that the neural circuitry for face processing consists of a network of brain regions in the temporal and prefrontal cortex as well as in the amygdala. During the past year, we have made a number of important discoveries regarding the neural mechanisms mediating face processing: 1. Monkeys with bilateral amygdala lesions, unlike normal monkeys, showed no preference for viewing faces or illusory faces relative to objects, indicating that the amygdala is critical for our earliest specialized response to faces, a behavior thought to be a precursor for efficient social communication. 2. Intranasal administration of oxytocin, a neuropeptide and potential treatment for autism, selectively altered fMRI responses to emotional expressions in monkeys, significantly reducing responses to both fearful and aggressive faces in face-responsive regions while leaving responses to appeasing and neutral faces unchanged, thereby explaining the beneficial effects of oxytocin on social cognition. 3. Unlike face-selective regions in the ventral stream, which all showed a contralateral visual field bias, that in the posterior superior temporal sulcus (pSTS) did not, indicating that this area may form part of a anatomically distinct visual pathway that preferentially processes moving biological information such as facial expressions and human movement. 4. Category-selective areas (for faces, bodies and scenes) located laterally in the human cortex were activated more by moving than by static visual stimuli (regardless of the object category), whereas those located ventrally were not, suggesting a new organizing principle for cortical areas in the human brain. 5. The greater activation evoked by fearful compared to neutral faces (termed the valence effect) was found throughout the entire visual field representation within visual cortical areas (that is, not confined to the retinotopic location of the faces), a finding consistent with the amygdala being the source of the valence effect inasmuch as the amygdala sends widespread projections back to visual cortex. 6. Patients with Moebius syndrome (MoS), a rare genetic neurological disorder characterized by facial paralysis, showed an impaired ability to recognize emotional expressions in others, but not in their ability to recognize expressions produced by body movements, indicating that the perception of facial emotional expressions depends on the ability to produce those expressions oneself. 7. Resting-state fMRI showed reduced posterior STS-amygdala connectivity relative to fusiform face area-amygdala connectivity in MoS compared to healthy controls, suggesting that the neurocircuitry underlying facial emotional expression processing may be altered in MoS. 8. Granger causality analysis of MEG data revealed that, during configural face processing but not during featural face processing, dorsal stream visual areas send information to ventral stream visual areas, indicating that visuospatial information is transmitted to face-processing regions during configural face processing. 9. As in monkeys, several curvature-selective regions of the posterior cerebral cortex were identified in humans that overlapped face-selective regions, suggesting that face selectivity might arise from curvature selectivity.