Research ReportFeatural and configural face processing differentially modulate ERP components
Introduction
Since all faces share the same first-order configuration (two eyes above a nose above a mouth), the discrimination and identification of individual faces is based on the processing of individual features (shape of the eyes, nose and mouth) and on the processing of the second-order configuration of these features (the distance between the eyes, or between the mouth and nose, etc.) (Maurer et al., 2002). Understanding how the event-related potentials (ERPs) are influenced by subtle differences in face features or configuration could help understand the typical time course of face processing, and especially how individual faces are discriminated and identified.
The P1 component is usually thought to be mainly influenced by basic visual parameters such as luminance and contrasts, by the direction of spatial attention or by the participant's state of arousal (Luck, 2005). However, there is also evidence that this early component is sensitive to violations of the first-order configuration of a face. For example, Boutsen et al. (2006) observed that ‘thatcherization’ (inversion of the eyes and mouth relative to the rest of the face) increased the amplitude of the P1 component. Macchi Cassia et al. (2006) scrambled faces along two spatial dimensions: vertical symmetry (the symmetry of the left and right halves of a face along a vertical axis) and up–down featural arrangement (the proportion of high-contrast features in the upper compared to the lower part of the face). They observed that asymmetrical faces elicited a larger P1 compared to symmetrical faces, and that the P1 was larger when the eyes were presented in the bottom half of faces (bottom-heavy) than when they were presented in the top half of faces (top-heavy). However, are these early processes affected when faces retain their natural first-order configuration? Is the P1 sensitive to these more subtle differences of the second-order configuration or of the features that make discrimination of individual faces possible? Halit et al. (2000) found that elongating faces by increasing the eyes–nose–mouth distance influenced the P1 amplitude. Moreover, the P1 was larger for natural faces judged as atypical relative to natural faces judged as more typical. In sum, the literature suggests that the P1 may index functions beyond the low-level processing of visual stimuli and might be sensitive to the first- and second-order configuration of a face. It is not clear if this component is also affected by modifications of face features.
It is well-established that the N170 component is modulated by face processing. Specifically, this component is larger when elicited by human faces than by other object categories (Bentin et al., 1996). The N170 is not affected by stimulus familiarity, and is therefore more likely to reflect structural encoding of face stimuli rather than identity recognition (Bentin and Deouell, 2000, Eimer, 2000). However, it remains unclear what exactly is encoded at the stage of the N170. There is some evidence that the N170 could be sensitive to the processing of the eye features in a face. Using the ‘Bubbles technique’, Schyns et al. (2003) observed that a larger N170 amplitude was associated with the availability of eye information. Moreover, Bentin et al. (2002) observed that pairs of abstract patterns elicited an N170 similar to that of faces when they were perceived as eyes, but an N170 similar to that of objects when they were not perceived as eyes. Previous literature also suggests that the N170 component is sensitive to the first-order configuration of a face. Indeed, Boutsen et al. (2006) observed a reduction of the N170 amplitude for ‘thatcherized’ faces compared to natural faces. The N170 was also observed to be of larger amplitude for asymmetrical than symmetrical faces and for top-heavy faces than for bottom-heavy faces (Macchi Cassia et al., 2006). With regard to second-order configuration, Halit et al. (2000) observed that elongation of faces did not influence the N170, whereas this component was larger for faces judged as atypical than for faces judged as typical. Moreover, Scott and Nelson (2006) observed that lateralization of the N170 component was influenced by featural and second-order configural differences in faces, although they found that the general amplitude of the N170 was not influenced by these differences in faces. When a difference waveform was calculated by subtracting the N170 response to the manipulated face from that of the familiarized face, the amplitude of this difference was significantly larger in the left hemisphere for featural than for configural manipulations, whereas the opposite was observed in the right hemisphere. Thus, the available literature suggests that the N170 ERP component could be sensitive to the first-order configuration of a face, as well as to some aspects of its second-order configuration and to the eye features, which is congruent with the idea that this component reflects the structural encoding stage of face processing.
The P2 component has a similar topography as the P1 component (Curran et al., 1993, Tucker et al., 1994), and may reflect the re-activation of the primary and secondary visual areas in a process of visual cortical feedback (Kotsoni et al., 2006, Kotsoni et al., 2007). The literature reveals modulations of the P2 component by modifications of the first-order configuration (‘thatcherization’; Boutsen et al., 2006) and second-order configuration (elongation; Halit et al., 2000) of a face, but it is not clear if this component is also sensitive to featural processing of faces.
The first aim of this experiment was to explore the brain processes involved in the discrimination of individual faces by systematically investigating the influence of feature and second-order configuration modifications of face stimuli on ERP components. In the first part of the experiment, participants were asked to judge whether pairs of faces that sometimes differed on their features or second-order configuration were the same or different. This task does not explicitly direct attention to either aspect of face processing, although the modifications of the face stimuli may implicitly direct attention to the features or configuration of these stimuli. The second aim of this experiment was to determine whether explicitly instructing participants to attend to the features or to the configuration of faces would influence the neural processes involved for these stimuli. By varying the experimental task to direct attention to the features of a face, it was expected that the whole face would be processed with a more local strategy, whereas directing attention to the configuration of a face would elicit a global processing strategy. In analogy with the influence of local/global processing strategies on the neural processes involved for hierarchical stimuli1 (Martínez et al., 1997, Moses et al., 2002), we hypothesized that some face-sensitive ERP components would be modified by tasks aimed at shifting attentional focus to features or configuration. To test this hypothesis, the second part of the experiment presented pairs of faces that always differed in their features or in their second-order configuration. Participants were asked to attend to the features or to the configuration in these faces and to decide if there was a difference in the attended aspect while ignoring other differences. This task is similar to the one used with hierarchical stimuli by Moses et al. (2002) and Martínez et al. (1997), where participants were asked to attend to either the local or global level of hierarchical stimuli and decide if a target shape or letter was present at the attended level. The stimuli used for this experiment were the “Jane stimuli” (Mondloch et al., 2002). We analyzed the impact of these tasks and stimulus conditions on the amplitude of the P1, N170 and P2 components. The P300 component was also studied to find if the two tasks were successful in varying the processing strategies. Since this component is generally larger for target stimuli (Luck, 2005), we predicted that the P300 would be larger for faces with a feature modification in the featural task and larger for faces with a configuration modification in the configural task if the tasks successfully influenced the participant’s focus of attention.
Section snippets
Behavioral results
A one-way ANOVA revealed that the accuracy was influenced by the task. Participants were more accurate in the featural task than in the same–different task or the configural task (see Fig. 1 and Table 1). A one-way ANOVA in the same–different task revealed that the accuracy differed across stimulus category, with more errors to Configuration and Original stimuli than to Feature stimuli. In the featural task there was no significant difference between the two types of stimuli, whereas in the
Discussion
This experiment studied the ERP correlates of featural and configural face processing. Part 1 explored how features and second-order configuration modifications in faces affect early ERP components. Part 2 tested the hypothesis that ERP correlates of faces could be modulated by instructing participants to attend to the features or the configuration of face stimuli.
The behavioral results from this study were congruent with previous results using the same (LeGrand et al., 2006, Maurer et al., 2007
Participants
13 participants between 22 and 38 years old (mean age = 28; 9 females) were paid for their participation. These participants were right-handed according to a modification of the Edinburgh Handedness Inventory and had normal or corrected-to-normal vision. They reported no language, reading or learning disorders, and none reported taking any psychoactive drug. The experiment was undertaken with the understanding and written consent of each participant.
Stimuli
The stimuli used for this experiment were the
Acknowledgments
We would like to thank Dr Cathy Mondloch and her colleagues for sharing the Jane stimuli with us.
This work was supported by MRC grant PG97 15587 to MJ and G0400341 to FD, as well as by the Canadian Institutes of Health Research and Birkbeck, University of London to EM.
References (25)
- et al.
Comparing neural correlates of configural processing in faces and objects: an ERP study of the Thatcher illusion
NeuroImage
(2006) - et al.
Topography of the N400: brain electrical activity reflecting semantic expectancy
Electroencephalogr. Neurophysiol.
(1993) Event-related brain potentials distinguish processing stages involved in face perception and recognition
Clin. Neurophysiol.
(2000)- et al.
Electrophysiological correlates of common-onset visual masking
Neuropsychologia
(2007) - et al.
What aspects of face processing are impaired in developmental prosopagnosia?
Brain Cogn.
(2006) - et al.
The many faces of configural processing
Trends Cogn. Sci.
(2002) - et al.
Neural correlates of processing facial identity based on features versus their spacing
Neuropsychologia
(2007) - et al.
Functional MRI of global and local processing in children
NeuroImage
(2002) - et al.
A holistic account of the own-race effect in face recognition: evidence from a cross-cultural study
Cognition
(2004) Spatial sampling of head electrical fields: the geodesic sensor net
Electroencephalography and Clinical Neurophysiology
(1993)