Summary: A Harvard Medical School study finds that exposure to faces during early development is essential for the brain regions that recognize faces to form. Monkeys raised without seeing faces failed to develop face-selective brain patches, showing that experience—rather than an innate template—drives face-domain formation.
Source: Harvard Medical School
Early Visual Experience Shapes the Brain’s Ability to Recognize Faces
For decades, scientists have assumed that the ability to recognize faces is innate in primates—that the brain arrives wired to detect and distinguish faces from birth. A new study from Harvard Medical School, published in Nature Neuroscience, challenges that idea. The research shows that specific brain regions specialized for face recognition develop only when an animal is exposed to faces during the early months of life.
Study Overview
Neurobiologists Margaret Livingstone, Michael Arcaro, and Peter Schade led experiments with rhesus macaques to test whether face-recognition areas in the brain require visual experience to form. Macaques are a common model for human brain development because of their close evolutionary relationship and similar neural organization. In typical macaque development, clusters of neurons that respond preferentially to faces—often called face patches—emerge in the superior temporal sulcus by around 200 days of age.
Experimental Design
The researchers raised two groups of infant macaques. The control group experienced normal social exposure: time with their mothers, with juvenile peers, and with human caretakers whose faces were visible. The experimental group was raised by human caregivers who bottle-fed, played with, and cuddled the infants while wearing welding masks. As a result, these face-deprived macaques did not see human or conspecific faces during the first year of life. After the deprivation period, all animals were placed in social groups where they could see faces.
At about 200 days old, the team used functional MRI to examine the visual cortex of both groups. They measured responses to multiple categories of visual stimuli, including faces, hands, objects, scenes, and bodies, and also tracked gaze behavior while the animals viewed images.
Key Findings
The control macaques showed consistent, category-selective regions for faces, hands, objects, scenes, and bodies, matching expectations from prior work. In contrast, the face-deprived macaques developed regions selective for hands, objects, scenes, and bodies, but they did not develop face-selective patches. Retinotopic organization across the visual system remained normal, indicating that the deprivation produced a highly selective deficit limited to face processing.
Behavioral observations matched the imaging results. Control animals naturally looked more at faces in images, even before face patches typically emerge. Face-deprived macaques did not show a preference for faces; instead, they preferentially looked at hands. Correspondingly, the brain territory responsive to hands was expanded in the deprived animals compared with controls.
Interpretation: Experience Drives Domain Formation
These results support a model in which early viewing behavior and environmental importance guide the development of category-selective visual domains. The authors propose a three-step process: environmental relevance biases what an infant looks at; those viewing patterns drive neuronal activity; and ongoing activity sculpts the formation of category-selective cortical domains in stereotyped locations of inferotemporal cortex. Because a retinotopic map exists at birth, preferential viewing of certain stimuli in particular parts of the visual field can bias where domains emerge without requiring innate, category-specific templates.
Implications for Developmental Disorders
The study sheds light on conditions in which face recognition is impaired. Developmental prosopagnosia, for example, involves lifelong difficulty recognizing familiar faces and may reflect a failure of face-domain formation. The findings also bear on autism spectrum disorders, where reduced attention to faces during early development is common. If lack of face exposure prevents normal formation of face-selective regions, early interventions that encourage looking at faces could help mitigate some social and perceptual deficits.
Authors, Funding and Methods
The research team included Margaret S. Livingstone, Michael J. Arcaro, Peter F. Schade, Justin L. Vincent and Carlos R. Ponce. Co-investigators included Justin Vincent and Carlos Ponce of Harvard Medical School. The work was supported by National Institutes of Health grants, a William Randolph Hearst Fellowship, and used resources at the Athinoula A. Martinos Center for Biomedical Imaging at Massachusetts General Hospital.
Conclusion
This study provides strong evidence that face recognition in primates is not an entirely innate ability but instead depends on early visual experience. What infants look at helps determine which neural circuits strengthen and which category-selective regions emerge. The findings emphasize the importance of early sensory and social environments in shaping perceptual and cognitive development and suggest practical avenues for interventions in developmental disorders associated with reduced face-looking.
Abstract (summary)
Monkeys raised without exposure to faces did not develop face-selective cortical domains but did develop domains for other categories and showed normal retinotopic organization. These results indicate that early visual experience is necessary for face-domain formation. Behavioral gaze tracking showed that control monkeys preferentially looked at faces even before face domains appear, whereas face-deprived monkeys did not, suggesting that face-looking is not innate. A preexisting retinotopic map may allow selective early viewing to bias category-specific responses toward particular retinotopic representations, leading to stereotyped domain formation without built-in category templates. Environmental importance influences viewing behavior, viewing behavior drives neuronal activity, and neuronal activity sculpts domain formation.
Published: Nature Neuroscience (study published online September 4, 2017)