Face processing specificity in 3 month old Infants

*Boston Children's Hospital- Division of Developmental Medicine, Laboratories of Cognitive Neuroscience
*Universiti degli Studi di Milano - Bicocca, Italy

A number of behavioral studies have shown that newborns' visual attention is spontaneously attracted toward face-like visual configurations, raising the possibility that an experience-independent mechanism dedicated to face processing exists at birth. Recent studies question this conclusion, suggesting that the apparent preference for faces over non-faces in newborns reflects non-specific basic visual preferences rather than content-determined preferences. For example, Turati et al. (2002) recently demonstrated that one of the general structural properties that is crucial in determining newborns' face preference is the presence of an up-down asymmetry in the distribution of the elements within the shape, with a higher density of elements appearing in the upper visual field. This evidence suggests that the face processing specificity, assumed in the adult perceptual and neural system, may be induced by an experience-expectant developmental process (Nelson, 2001), through which circuits in the developing cortical pathways involved in visual recognition gradually become specialized for face processing.

Our hypothesis is that selectivity for the specific geometry of the face may be learned during the first 2 or 3 months of life arising from the non-specific attentional bias toward patterns with more features in the upper versus lower half. In the current study we predicted that selectivity for the specific geometry of the face may emerge during the first 3 months of life as a product of perceptual narrowing resulting from experiential input, leading to the construction of the first broadly defined face category segregating faces from other visual objects which may share with faces one or more visual properties. In the current investigation we examined whether there is electrophysiological differentiation at 3 months between face and non-face stimuli. Event-related-potentials (ERPs) were recorded from 62 scalp electrodes in three separate experiments: 1) while infants viewed pictures of faces and scrambled faces which were matched to the natural face for the number of elements appearing in the upper part of the configuration, 2) while infants viewed pictures of faces and geometric patterns in which all characteristics common to faces were removed, except the top-heavy property, and 3) while infants viewed pictures of faces, as well as toys that shared no specific properties with faces.

ERP results from Experiment 1 showed evidence of differentiation between the face and scrambled stimuli, but only for the N700 component. No differentiation was found for earlier components that are thought to reflect the adult-like structural encoding stage of face processing in infants (N290 and P400). Results from Experiment 2 also revealed no differentiation between faces and top-heavy geometric stimuli for early ERP components (N290, P400). Experiment 3 revealed differentiation between faces and toys for the P400. All categories of non-face stimuli were differentiated from faces at later attentional components. These findings suggest that the top-heavy property still plays a role in triggering brain responses to faces at 3 months, but by this time faces have been salient enough to recruit differential attentional responses. This may reflect the high degree of familiarity and/or the social value faces have gained over the infants' first 3 months of life.

Implications of this work: There are a number of childhood disorders that result in facial processing deficits. This project was designed to more closely examine aspects that drive infants' basic visual preferences for faces early in life. By better understanding how a specific face template develops during the first 2-3 months of life, we may be better able to identify infants at risk for developing difficulties with facial processing. Early identification of such individuals would allow us to develop more effective and time-sensitive intervenions.