Face processing as a brain adaptation at multiple timescales

Abstract

I consider face processing as the brain's adaptive response to phylogenetic, ontogenetic, and task-specific factors. Focusing on wide-ranging evidence from both my own laboratory and others, evidence for a primitive “quick and dirty” route for face processing that exists prior to postnatal experience is reviewed. Next, I trace the emergence of cortical specialization for face processing influenced by individual developmental experience (ontogenetic adaptation) and suggest that this ontogenetic adaptation is also heavily constrained by the phylogenetic system. Finally, I turn to recent evidence on task-specific modulation of activity in the core face network that illustrates brain adaptation at a finer timescale than that for the other systems. Current evidence indicates that task-specific modulation of the cortical face network does not emerge until the teenage years. As previously proposed for other components of cognition, I propose that these systems are complementary to each other, each compensating for the others' weaknesses. Different face-related systems are adapted to respond to survival pressures at different timescales, from millennia, to months, to microseconds.

Keywords:

• Face perception
• Brain development

Acknowledgments

I wish to thank my many colleagues and collaborators who have contributed to my thinking and research on face processing over the past decades. I am also grateful to the UK Medical Research Council for long-term funding, currently G0701484.

Notes

¹ The proposal of a “low road” or “quick and dirty” route has recently been brought into question in the context of studies of emotion (Pessoa & Adolphs, 2010). While some of the arguments put forward by these authors are specific to the processing of emotion, rather than faces, others are directly relevant to the evidence presented here. Pessoa and Adolphs proposed that rather than a separate subcortical route that directs cortical activity, it is more useful to consider waves of neural responses to biologically relevant stimuli in which even the most rapid initial response may involve both cortical and subcortical activity. The extent to which this alternative view is applicable to human newborns with a relative immature cortex remains to be determined.

² Not all developmental fMRI studies have reported changes in the fusiform face area (FFA) during development as described in the main text. For example, Pelphrey, Lopez, and Morris (2009) report adult-like responses in FFA, extrastriate body area (EBA), and the parahippocampal place area from ages 7 to 11 years. However, while an FFA was detectable in children, and voxels in that region showed the same degree of specificity as that seen in adults, many fewer voxels (around half) were classified as showing a face-selective response in children than in adults (see also Cantlon, Pinel, Dehaene, & Pelphrey, 2011). Thus, in accord with the results described in the main text, some evidence for an increase in the extent of face-selective tissue was provided. One currently unpublished study cited in secondary sources (Kanwisher, 2010) suggests that there may be no developmental change in the FFA when children from 5 years old are scanned using a number of different data acquisition and analysis methods that improve signal-to-noise ratio. However, it remains puzzling why neighbouring object-sensitive regions did not show similar developmental changes in all the previous studies—which also used a variety of different scanners, headcoils, and analytical methods.