Psychophysiology, 53 (2016), 1165–1173. Wiley Periodicals, Inc. Printed in the USA. C 2016 The Authors. Psychophysiology published by Wiley Periodicals, Inc. on behalf of Society for Psychophysiological Research V
DOI: 10.1111/psyp.12669
Double dissociation of configural and featural face processing on P1 and P2 components as a function of spatial attention
HAILING WANG, SHICHUN GUO,
AND
SHIMIN FU
Department of Psychology, Tsinghua University, Beijing, China
Abstract Face recognition relies on both configural and featural processing. Previous research has shown that P1 is sensitive to configural face processing, but it is unclear whether any component is sensitive to featural face processing; moreover, if there is such a component, its temporal sequence relative to P1 is unknown. Thus, to avoid confounding physical stimuli differences between configural and featural face processing on ERP components, a spatial attention paradigm was employed by instructing participants to attend an image stream (faces and houses) or an alphanumeric character stream. The interaction between attention and face processing type on P1 and P2 components indicates different mechanisms of configural and featural face processing as a function of spatial attention. The steady-state visual evoked potential (SSVEP) results clearly demonstrated that participants could selectively attend to different streams of information. Importantly, configural face processing elicited a larger posterior P1 (approximately 128 ms) than featural face processing, whereas P2 (approximately 248 ms) was greater for featural than for configural face processing under attended condition. The interaction between attention and face processing type (configural vs. featural) on P1 and P2 components indicates that there are different mechanisms of configural and featural face processing operating as functions of spatial attention. While the P1 result confirms previous findings separating configural and featural face processing, the newly observed P2 finding in the present study extends this separation to a double dissociation. Therefore, configural and featural face processing are modulated differently by spatial attention, and configural face processing precedes featural face processing. Descriptors: Face, Configural processing, Featural processing, Spatial attention, Event-related potentials (ERPs), Steady-state visual evoked potential (SSVEP) Evidence that different mechanisms are at work in configural (second-order relations) relative to in featural face processing is obtained from multiple sources, including behavioral studies (Freire, Lee, & Symons, 2000; Leder & Carbon, 2006), electrophysiological recordings (Mercure, Dick, & Johnson, 2008; Scott & Nelson, 2006; Wang, Sun, Ip, Zhao, & Fu, 2015), neuroimaging (Maurer et al., 2007; Renzi et al., 2013; Rotshtein, Geng, Driver, & Dolan, 2007), and prosopagnosia patients (Le Grand et al., 2006). However, ERP studies have resulted in conflicting findings on the N170 component. While the hemisphere lateralization of the N170 component was found to be influenced by configural and featural face processing (Scott & Nelson, 2006), other studies have not found any differences between configural and featural face processing regarding N170 (Mercure et al., 2008; Wang et al., 2015). Moreover, previous studies have shown that P1 is sensitive to configural face processing (Halit, de Haan, & Johnson, 2000; Wang et al., 2015) and that configural face processing enhances the amplitude of P2 relative to featural face processing (Halit et al., 2000; Mercure et al., 2008). These results suggest that the P1, N170, and P2 components are potentially sensitive to differentiating between the temporal order and the brain mechanisms involved in configural and featural face processing. Directly comparing ERPs elicited by stimuli with physical differences would violate the Hillyard Principle, which requires
Humans have a remarkable ability to rapidly detect faces among nonface stimuli and to discriminate or recognize thousands of individual faces. This ability relies on several types of face processing, which include the following: (a) first-order relations (i.e., the basic arrangement of each face with two eyes above a nose that itself is above a mouth), which allow individuals to quickly detect faces among a variety of stimuli; (b) second-order relations (e.g., the distance between the eyes or between the eyes and the mouth), also known as configural face processing, which allow one to recognize individual faces; and (c) featural processing, which allows faces to be individuated based on the shape of internal features (such as the eyes and mouth; Maurer, Le Grand, & Mondloch, 2002).
This research is supported by the National Natural Science Foundation of China (31371142) and the Initiative Scientific Research Program, Tsinghua University, China (2011Z02177) to SF. We thank Wenfeng Feng, Ruyi Qiu, and Chenteng Ip for their assistances with programming, data collection, and language modification, respectively. Address correspondence to: Dr. Shimin Fu, Department of Psychology, Tsinghua University, Haidian District, Beijing, China, 100084. E-mail:
[email protected];
[email protected] This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes. 1165
H. Wang, S. Guo, and S. Fu
1166 comparisons of ERPs be elicited by the same physical stimuli and allows for variances only with respect to psychological conditions (Luck, 2005). Thus, tricks to avoid such confounding of physical stimuli have been used in previous studies, including superimposed images (Sreenivasan, Goldstein, Lustig, Rivas, & Jha, 2009), attention (Wang et al., 2015), and adaptation paradigms (Feng, Luo, & Fu, 2013; Fu, Feng, Guo, Luo, & Parasuraman, 2012). The intent is either to keep the stimuli the same (e.g., the superimposed images) or to balance the physical differences between stimuli by comparing the attention or adaptation effects of two different stimuli. The balancing procedure was used in our previous study (Wang et al., 2015), in which attention was manipulated by stimulus onset asynchrony in a rapid serial visual presentation (RSVP) paradigm. Employing the same logic of balancing physical differences between configural and featural faces, a spatial attention paradigm was employed in the current study, in which participants were instructed to attend to a face/house stream or to an alphanumeric character stream. We proposed that a main effect of face processing type (i.e., configural vs. featural) would suffer from only the confounding of differences in physical stimulus and that a significant interaction between attention and face processing type would exclude that possibility because the physical differences between face processing types was balanced between the attended and unattended conditions. In comparison with the face recognition task of the RSVP paradigm in our previous study (Wang et al., 2015), three critical modifications should be noted. First, the incongruent task difficulty issue between configural and featural face processing in our previous study was solved by designing a task that required the detection of stimuli (the blurred images) that are irrelevant to configural and featural face processing, such that configural and featural processing were examined in a more implicit and task-irrelevant way in the present study. Second, in the present study, configural and featural face stimuli did not involve the working memory process, which contrasts with the face recognition task in the attentional blink that might be influenced by working memory encoding (Dux & Marois, 2009). Third, the RSVP of the alphanumeric character stream might result in a better manipulation of the unattended face condition (Crist, Wu, Karp, & Woldorff, 2008; Feng, Martinez, Pitts, Luo, & Hillyard, 2012) than our previous unattended face condition during the attentional blink. In short, the present design enabled us to assess the mechanism of configural and featural face processing as a function of spatial attention and permitted better manipulation of attention without confounding task and working memory. In brief, as all faces share the same first-order relations, individual faces are recognized through featural processing (the shape of internal features) and by second-order configural processing (the distance of the eyes or the mouth from the nose). Understanding how the ERP components respond to configural and featural face processing without being contaminated by physical differences may facilitate the understanding of the time course of individual face processing. Thus, the goal of the present study was to resolve the issue of physical difference by examining the brain process of second-order configural and featural face processing on ERP components as a function of spatial attention. We proposed three predictions regarding the results of the present study. First, as with our previous study (Wang et al., 2015), there should be a significant interaction between attention and face processing type (configural and featural face processing) on P1, which suggests that configural and featural face processing modulate ERP
Figure 1. Illustrations of the stimuli and procedures employed in the experiment. A: Illustrations of configural face processing differed in the distance between the eyes or between the mouth and nose. Illustrations of featural face processing differed in the shape of the eyes or month. B: Procedure. A series of face and house stimuli were presented in the left visual field. An RSVP stream of alphanumeric characters was presented in the right visual field. The location of face/house stimuli and alphanumeric characters was balanced. Participants were instructed to attend the left or right visual field in one block and pressed the button when the targets presented. C: Examples of standard stimuli and target stimuli. Digit targets were approximately 2% of the character stream. Blurred image targets comprised 20% of the face/house stream.
components differently. Second, and also similar to our previous study, N170 should reveal no difference between configural and featural face processing. Finally, as configural face processing was expected to elicit an earlier ERP response than featural processing on the P1 component—and as N170 was expected to be insensitive to configural and featural face processing—a later component, such as P2, should serve as the index for featural face processing. Method Participants Twenty students (12 females, 8 males, Mage 5 20.6 years, age range: 18–24 years) were recruited from Tsinghua University. Two students were excluded due to bad EEG data. Participants were paid for their participation, and all were healthy and right-handed with normal or corrected-to-normal vision. The research protocol was approved by the local Institutional Review Board at the Department of Psychology, Tsinghua
Double dissociation of configural and featural face processing
1167
University. Written informed consent forms were collected for all participants prior to the experiments.
Figure 1C). The blurry target images were equally likely to be either a face or a house.
Stimuli and Apparatus
Data Recording
Face stimuli were frontal view photos generated by FaceGen Modeller 3.5 (Toronto, Canada). The faces were hairless, had neutral expressions, did not feature glasses, and did not show their teeth. Configural and featural faces were the same as in our previous research (Wang et al., 2015). Configural face processing was constructed such that the distance between the eyes was either close or far (moving the eye position by 7 pixels [0.258] inward or outward), and the mouth was either close or far from the nose (moving the mouth position by 7 pixels [0.258] upward or downward; see Figure 1A). In the featural face processing set, the eyes and the mouth were simply replaced with different eyes and mouth (see Figure 1A). Overall, there were 12 pictures of configural face processing and 12 different featural faces. Pictures of 12 different houses were sourced from the Internet (http://image.baidu.com). All the stimuli were grayscale photographs of the same size (48 3 5.58), background, luminance, and other physical properties. Each picture was cropped using Adobe Photoshop CS5 software (San Jose, CA). All stimuli were presented on a 17-inch ViewSonic monitor (resolution: 1,024 3 768; refresh rate: 100 Hz) using MATLAB (R2010b, The MathWorks, Natick, MA). The viewing distance was 57 cm.
Brain electrical activity was recorded from 64 scalp sites with Ag/ AgCl electrodes mounted on an elastic cap (NeuroScan, TX), with the physical reference electrode situated on the Ref electrode site, which is located between CPz and Cz. Horizontal electrooculograms (EOGs) were recorded from two electrode sites at the outer canthi of each eye. Vertical EOGs were recorded from electrodes situated on infraorbital and supraorbital regions of the left eye. The interelectrode impedance was maintained below 5 kX throughout the EEG recording session. EEG and EOG readings were collected with a band-pass of 0.05–100 Hz and sampled at a rate of 500 Hz.
Procedure Participants were seated comfortably in a reclining chair and were trained to maintain fixation on the central cross. The procedure was similar to that in previous studies (Crist et al., 2008; Feng et al., 2012). Two streams of stimuli were presented concurrently, one in the left and one in the right visual field (see Figure 1B). One stream consisted of an RSVP of alphanumeric characters (28 3 28 visual angle), centered approximately 58 to the central fixation cross. Each character appeared for 150 ms before being replaced by a subsequent character, resulting in a stimulation rate of 6.67/s. The other stream of stimuli consisted of images of configural faces, featural faces, and houses (both 48 3 5.58 visual angle), which were presented in a random order at an eccentricity of 68 from the center of fixation to the center of the image. Each image was presented for 100 ms, and the intervals between image onsets were randomly varied between 600 and 900 ms. Each participant performed 16 blocks, each consisting of 136 face/house images. Before beginning each block, participants were instructed to attend to either the character stream or the face/house stream and to indicate the appearance of an occasional target in the designated stream by pressing a key. In one half of the blocks, the face/house stream was presented on the right side, and the character stream was presented on the left. In the other half of the blocks, the left/right locations were reversed. The blocks of the face/house and character streams were counterbalanced using different ABAB, BABA, ABBA, or BAAB series across different participants. The locations of stimuli were counterbalanced with respect to whether they were attended or unattended. When attending to the character stream, participants attempted to detect the appearance of infrequently presented digits (approximately 2% of the alphanumeric character stream) among (mostly uppercase) alphabetical characters. Targets in the face/house stream were blurred images of these objects and made up 20% of the number of images presented (see
Data Measurements and Analysis Electrophysiological data were analyzed using the EEGLAB toolbox in MATLAB (R2010b, The MathWorks). The EEG analyzing window was between 2200 and 600 ms, and the 200-ms prestimulus EEG served as a baseline. EEG data were band-pass filtered with a range of 0.1 to 30 Hz and rereferenced to the average of all electrodes. Artifact rejection was performed for all the EEG channels, and the rejection criterion was 6 75 mV. To evaluate the effect of attention, the steady-state visual evoked potential (SSVEP) elicited by the RSVP stream of characters was recorded and quantified. These SSVEP epochs were averaged for each subject and for the two attention conditions separately in the time domain from 2200 to 698 ms epochs with a sampling rate of 500 Hz. Thus, the window size was 450, with an integer number of cycles. The SSVEP amplitudes were calculated from the time domain averages using discrete Fourier transform and were quantified as the absolute value of the complex Fourier coefficients at 6.67 Hz. The SSVEP amplitudes were measured and averaged over a cluster of eight posterior electrode sites (P7/P8, PO3/PO4, PO7/PO8, O1/O2) and then tested by analysis of variance (ANOVA) using the factors for attention (attended vs. unattended), hemisphere, and electrode sites. In this study, the anterior N1, posterior P1, N170/VPP (vertex positive potential), and P2 components evoked by the faces/houses sequence were analyzed with the peak latencies (from the onset of the stimulus to the peak of each component) and amplitudes (baseline to peak). F1/F2, F3/F4, FC1/FC2, FC3/FC4, C1/C2, C3/C4 were selected for statistical analysis of the anterior N1 component (90–190 ms); P3/P4, P5/P6, P7/P8, PO3/PO4, PO5/PO6, PO7/PO8, O1/O2 were selected for statistical analysis of the posterior P1 component (80–180 ms) and P2 component (180–300 ms); N170 (150–270 ms) was analyzed at the P7/P8, PO7/PO8 electrode sites; and VPP (150–270 ms) were analyzed at the Fz/FCz/Cz/CPz electrode sites. Analyses were performed on data collapsed across both the left and right visual fields of stimuli, as this factor was irrelevant to our critical hypotheses and tests. To assess the effects of attention on face processing type, a four-way repeated measures ANOVA was performed on the amplitude and latency of each component using the factors of attention (attended vs. unattended), face processing type (configural vs. featural), hemisphere, and electrode sites as within-subject factors. In addition, to assess the effects of attention on face and object processing, a four-way repeated measures ANOVA on the amplitude and latency of N170/VPP was conducted using the factors of attention (attended vs. unattended), object type (face vs. house), hemisphere, and electrode sites as
H. Wang, S. Guo, and S. Fu
1168
Figure 2. A: Grand-averaged SSVEP amplitudes recorded at sites of PO7/PO8 demonstrating that the attended RSVP stream of characters had larger SSVEP amplitudes than the unattended condition. B: Topographical distributions of SSVEP amplitudes at 6.67 Hz to character stream are shown separately for the attended, unattended conditions and for the differences between the attended and unattended condition.
within-subject factors. When necessary, p values were corrected using Greenhouse–Geisser correction. In addition, Bonferroni correction was used for multiple comparisons and post hoc analyses. Results Behavioral Performance When participants attended to the alphanumeric stream, the mean reaction time (RT) for detecting the digit targets was 515 6 48 ms, and the mean hit rate was 84% 6 10.5%. When subjects attended to the face/house stream, the mean RT for detection of the blurry targets was 461 6 38 ms for the faces and 457 6 34 ms for the houses. The corresponding hit rates were 92.2% 6 10% for the blurry face targets and 95.1% 6 9.5% for the blurry house targets. Data analysis revealed that the hit rates of face/house targets were greater than the rates for the digit targets, F(1,17) 5 13.58, p < .002, gp2 5 .44. The face/house targets exhibited shorter RT than the digit targets, F(1,17) 5 45.87, p < .001, gp2 5 .73. No significant effects were found between the face and house targets. This result indicates that it is easier for participants to detect targets embedded in the face/house stream than in the character stream and that the unattended face condition was better controlled. Moreover, both performance levels were above chance, which indicated the effective allocation of attention to the assigned tasks. ERP Data Analysis of the SSVEP The SSVEP amplitude to the RSVP stream of characters was enlarged by attention (.95 vs. .22 mV), F(1,17) 5 27.82, p < .001,
gp2 5 .62, thus reflecting enhanced sensory processing of the RSVP stream when it was attended. No other significant effects were found (see Figure 2). ERP Data Analysis of the Effect of Spatial Attention on Face Processing Type Anterior N1. The anterior N1 amplitudes exhibited significant main effects for attention, F(1,17) 5 21.53, p < .001, gp2 5 .56. The attended face elicited much larger anterior N1 amplitudes than the unattended face (22.48 vs. 2.61 mV). There was no significant interaction between attention and face processing type, F(1,17) 5 1.39, p < .25, gp2 5 .08, or for Attention 3 Face Processing Type 3 Hemisphere 3 Electrode Sites, F(5,85) 5 .49, p < .65, gp2 5 .03. Posterior P1. The posterior P1 amplitudes revealed significant main effects for attention, face processing type, and hemisphere, F(1,17) 5 25.67, p < .001, gp2 5 .6; F(1,17) 5 4.71, p < .044, gp2 5 .22; F(1,17) 5 7.4, p < .015, gp2 5 .3, respectively. Attended face elicited much larger posterior P1 amplitudes than the unattended face (4.32 vs. 1.1 mV; see Figure 3). The posterior P1 amplitude was significantly larger for configural face processing than for featural face processing (2.79 vs. 2.63 mV), and more positive posterior P1 amplitudes were found in the right hemisphere than in the left hemisphere (3.06 vs. 2.36 mV). More importantly, there was a significant interaction between attention and face processing type, F(1,17) 5 3.52, p < .048, gp2 5 .17, which suggests that the differences in physical stimuli were manipulated. A post hoc analysis indicated that configural processing elicited a greater positive amplitude
Double dissociation of configural and featural face processing
1169 processing type, F(1,17) 5 3.21, p < .05, gp2 5 .16, suggests that the difference between configural and featural face processing was not confounded by physical stimuli. A post hoc analysis showed that featural processing elicited a greater positive amplitude than did configural processing when the face was attended (4.37 vs. 3.96 mV, p < .025). In contrast, the effect of face processing type was entirely absent when the face was unattended (see Figure 4 and 5B). ERP Data Analysis of the Effect of Spatial Attention on Object Type
Figure 3. Grand-averaged ERP waveforms for the attended and unattended condition recorded at sites of P7/P8 and PO7/PO8 on face (A) and house stimuli (B). The attended condition had larger P1 and N170 amplitudes than the unattended condition both on face and house stimuli.
than did featural processing when faces were attended (4.5 vs. 4.14 mV, p < .018). By contrast, there were no effects of face processing type when faces were unattended (see Figure 4 and 5A). Furthermore, the posterior P1 latency indicated significant main effect for attention, F(1,17) 5 7.73, p < .013, gp2 5 .31, and longer latencies elicited by the attended face than by the unattended face (132 vs. 124 ms). N170/VPP. The N170 amplitudes presented a significant main effect for attention, F(1,17) 5 7.43, p < .014, gp2 5 .3, and the N170 component was enhanced in trials in which the face was attended in comparison with trials in which the face was unattended (22.18 vs. 21.21 mV; see Figure 3). As with the N170 amplitudes, VPP amplitudes also revealed significant main effect for attention, F(1,17) 5 7.76, p < .013, gp2 5 .3, as the VPP component was enhanced in trials in which face was attended relative to trials in which the face was unattended (1.76 vs. .81 mV). P2. The P2 amplitudes also showed significant main effects for attention, face processing type, and hemisphere, F(1,17) 5 12.61, p < .002, gp2 5 .43; F(1,17) 5 5.13, p < .037, gp2 5 .23; F(1,17) 5 6.55, p < .02, gp2 5 .28, respectively. The attended face elicited much larger P2 amplitudes than unattended face (4.17 vs. 1.12 mV; see Figure 3). The P2 amplitude was significantly larger for featural face processing than for configural face processing (2.76 vs. 2.53 mV), and more positive P2 amplitudes were found in the right hemisphere than in the left hemisphere (2.89 vs. 2.41 mV). Moreover, the significant interaction between attention and face
The N170 amplitudes revealed a significant main effect for object type, F(1,17) 5 101.84, p < .001, gp2 5 .86. Faces elicited a greater negative N170 amplitudes than did houses (21.7 vs. 1.47 mV). Furthermore, there was a significant interaction between attention and object type, F(1,17) 5 25.47, p < .001, gp2 5 .6. A post hoc analysis indicated that faces elicited greater negative amplitudes when compared with houses under the attended condition (22.18 vs. 2.3 mV, p < .001) and that faces also elicited greater negative amplitude when compared with houses under the unattended condition (21.21 vs. .63 mV, p < .001; see Figure 6). The differences between the face- and house-evoked N170 were larger under attended condition than under unattended condition (24.49 vs. 21.84 mV, p < .001). VPP amplitudes showed a significant main effect of object type, F(1,17) 5 45.26, p < .001, gp2 5 .73, and faces elicited much larger VPP amplitudes than those elicited by houses (1.29 vs. 2.49 mV). There was significant interaction between attention and object type, F(1,17) 5 28.68, p < .001, gp2 5 .63. A post hoc analysis revealed that faces elicited larger amplitudes than houses elicited under the attended condition (1.76 vs. 2.79 mV, p < .001), and that faces also elicited greater positive amplitudes than houses under the unattended condition (.81 vs. 2.2 mV, p < .001). The differences between the face- and house-evoked VPPs were significantly larger in the attended condition than in the unattended condition (2.55 vs. 1.01 mV, p < .001). Discussion The central aim of the present study was to investigate the brain processes of configural and featural face processing on ERP components with spatial attention. In the present study, the effective manipulation of spatial attention is demonstrated by the results of SSVEP amplitudes and anterior N1 and posterior P1 components. Our results confirmed previous findings that SSVEP amplitudes are enhanced by the attended condition (Mishra, Zinni, Bavelier, & Hillyard, 2011; Morgan, Hansen, & Hillyard, 1996; M€uller & Hillyard, 2000), and they show that SSVEP is suitable for the study of spatial attention. Moreover, consistent with previous findings (Crist et al., 2008), the N170/VPP difference waves between face and house stimuli are enhanced under the attended condition relative to the unattended condition. Importantly, there is a significant interaction between face processing type and attention. In particular, the posterior P1 amplitude is greater for configural face processing than for featural face processing, whereas featural face processing elicits a larger P2 than configural face processing under the attended condition, which represents a novel finding of double dissociation between configural and featural face processing on the P1 and P2 components as functions of attention. Remarkably, in comparison with our previous study (Wang et al., 2015), this double dissociation is obtained in a design with a better manipulation of attention and without being contaminated by factors such as physical difference, task, and working memory.
1170
H. Wang, S. Guo, and S. Fu
Figure 4. Grand-averaged ERP waveforms of configural and featural face processing recorded at sites of P7/P8 and PO7/PO8 when they were attended (A) and unattended (B). Topographic maps of configural and featural face processing when attended at 128 ms and 248 ms (at the bottom of A). Configural face processing had larger P1 amplitudes than featural face processing, and featural face processing had larger P2 amplitudes than configural face processing under attended condition.
Configural and Featural Face Processing Differently Modulate the P1 and P2 Components The interaction between attention and face processing type (configural and featural face processing) is observed in both the P1 and P2 components. Consistent with our previous study (Wang et al., 2015), configural face processing has a larger posterior P1 than featural face processing under the attended condition. Moreover, a new finding of the present study is that featural face processing elicits a larger P2 relative to configural face processing under the attended condition. This result not only suggests that face processing type strongly depends on spatial attention, but also shows that configural and featural face processing are processed over different time courses under the attended condition. Two streams of evidence support the notion that P1 is sensitive to configural face processing. First, numerous studies of face perception have reported that inverted faces elicit larger P1 amplitudes
than upright faces (Itier & Taylor, 2002, 2004a) and that faces elicit larger P1 amplitudes than objects (Herrmann, Ehlis, Muehlberger, & Fallgatter, 2005). These results suggest that P1 might reflect the holistic processing of a face as a face (Itier & Taylor, 2004b). Second, previous research has shown that P1 was modulated by attention to motion but not by attention to color (Zanto & Gazzaley, 2009), implying its sensitivity to the processing dissociation between the dorsal (for motion and spatial relationship processing) and ventral (for color and shape processing) visual pathways (Ungerleider, Courtney, & Haxby, 1998; Van Essen & Gallant, 1994). This finding may help explain the increasing P1 component for configural relative to featural face processing, because configural face processing differentiates spacing (dorsal processing) of individual features, whereas featural face processing differentiates the shapes (ventral processing) of individual features. By contrast, the relationship between the P2 component and featural processing is less evident. As previous research has shown, the
Double dissociation of configural and featural face processing
1171
Figure 5. A: Mean amplitudes of P1 (P3/P4, P5/P6, P7/P8, PO3/PO4, PO5/PO6, PO7/PO8, O1/O2) elicited by configural and featural processing when they were attended and unattended. B: Same as A for the P2 amplitudes. *p < .05.
occipital P2 component is sensitive to face inversion, Thatcherization (a face with inverted eyes and mouth looks abnormal when upright but not when inverted), and contrast reversal (Boutsen, Humphreys, Praamstra, & Warbrick, 2006; Itier & Taylor, 2002), thus providing tentative evidence for the later stage of face processing. Importantly, as opposed to previous studies (Mercure et al., 2008), the results from the current study show a larger P2 for featural
face processing than for configural face processing. However, as participants’ task in the present study does not involve face processing type, the P2 component may reflect differences in configural and featural face processing rather than the deep processing of configural faces resulting from task difficulty, as discussed by Mercure et al. (2008). Further studies must be conducted to resolve the question of the sensitivity of the P2 component to featural face processing. Taken together, our results on the P1 and P2 components provide evidence of a double dissociation between the time course of configural and featural face processing under the attended condition. N170/VPP Is Sensitive to Object Type but Insensitive to Configural/Featural Face Processing
Figure 6. Grand-averaged ERP waveforms for faces and houses recorded at sites of P7/P8 and PO7/PO8 when they were attended (A) and unattended (B). The N170 difference waves between face and house stimuli were enhanced when they were attended compared to when they were unattended condition.
The N170/VPP components have been described in previous studies as face-sensitive responses (Bentin, Allison, Puce, Perez, & McCarthy, 1996; Eimer, 2000; Joyce & Rossion, 2005). Consistent with previous ERP research with the similar paradigm (Crist et al., 2008), our results also showed that the processing of faces was distinct from the processing of houses and that this categorical processing was modulated by spatial attention. However, Feng et al. (2012) did not find the effects of attention on this categorical processing in a similar paradigm. Differences in stimuli might explain the discrepancy. For example, both faces were upright in Crist et al.’s (2008) study and our study, whereas inverted faces that might have disrupted configural processing were employed in Feng et al.’s (2012) study. Meanwhile, the analyses of ERPs were performed only over contralateral electrode sites in Feng et al.’s (2012) study, which might also have led to the discrepancy. Moreover, consistent with our prediction and consistent with previous findings (Mercure et al., 2008; Wang et al., 2015), the N170/VPP component was not modulated by face processing type, regardless of whether faces were attended or unattended. Thus, the present result provided evidence that N170/VPP is sensitive to the firstorder relations of faces (Bentin et al., 1996; Itier, Latinus, & Taylor, 2006; Latinus & Taylor, 2006) but insensitive to the steps of recognizing individual faces, that is, the configural (second-order relations) and featural face processing.
H. Wang, S. Guo, and S. Fu
1172 Limitations Although P1 seems to be more sensitive to low-level stimulus properties (Rossion & Jacques, 2008), low-level differences cannot provide an adequate explanation for the effects found on P1 amplitude. As the featural faces used in the experiment look like normal and boring faces and the configural faces usually look a bit abnormal and novel because of the varied distance between facial features, the configural faces stimuli might attract more visual attention at early stage (i.e., posterior P1 component). To eliminate this possibility, we analyzed the frontocentral N2 and P3a, two indicators of attention orientation to the novelty (Folstein & Van Petten, 2008; Polich, 2007).1 No significant differences on these components were found with respect to the processing of configural and featural faces. Thus, these results showed that the slight novelty of configural faces stimuli is not the main cause of the differences involved in processing configural versus featural faces on posterior P1 components. In addi-
tion, it is further noted that the ERP responses to configural and featural face processing are distinct from the one to the global and local processing of hierarchical letters, which indicated that P1 and P2 were sensitive to local and global conditions, respectively (Han, Liu, Yund, & Woods, 2000). Thus, the psychological significance of the double dissociation between configural and featural face processing on P1 and P2 components merits further investigation. Conclusion P1 is more sensitive to configural than featural face processing, whereas P2 exhibits the opposite effect under the attended condition. This double dissociation between configural and featural face processing suggests that different underlying mechanisms were involved in these two steps of face recognition, and configural precedes featural face processing.
References Bentin, S., Allison, T., Puce, A., Perez, E., & McCarthy, G. (1996). Electrophysiological studies of face perception in humans. Journal of Cognitive Neuroscience, 8, 551–565. doi: 10.1162/jocn.1996.8.6.551 Boutsen, L., Humphreys, G. W., Praamstra, P., & Warbrick, T. (2006). Comparing neural correlates of configural processing in faces and objects: An ERP study of the Thatcher illusion. NeuroImage, 32, 352– 367. doi: 10.1016/j.neuroimage.2006.03.023 Crist, R. E., Wu, C. T., Karp, C., & Woldorff, M. G. (2008). Face processing is gated by visual spatial attention. Frontiers in Human Neuroscience, 1–6. doi: 10.3389/neuro.09.010.2007 Dux, P. E., & Marois, R. (2009). The attentional blink: A review of data and theory. Attention Perception & Psychophysics, 71(8), 1683–1700. doi: 10.3758/APP.71.8.1683 Eimer, M. (2000). The face-specific N170 component reflects late stages in the structural encoding of faces. NeuroReport, 11, 2319–2324. doi: 10.1097/00001756-200007140-00050 Feng, C., Luo, Y., & Fu, S. (2013). The category-sensitive and orientationsensitive N170 adaptation in faces revealed by comparison with Chinese characters. Psychophysiology, 50, 885–899. doi: 10.1111/psyp. 12067 Feng, W. F., Martinez, A., Pitts, M., Luo, Y. J., & Hillyard, S. A. (2012). Spatial attention modulates early face processing. Neuropsychologia, 50, 3461–3468. doi: 10.1016/j.neuropsychologia.2012.09.031 Folstein, J. R., & Van Petten, C. (2008). Influence of cognitive control and mismatch on the N2 component of the ERP: A review. Psychophysiology, 45, 152–170. doi: 10.1111/j.1469-8986.2007.00602.x Freire, A., Lee, K., & Symons, L. A. (2000). The face-inversion effect as a deficit in the encoding of configural information: Direct evidence. Perception, 29, 159–170. doi: 10.1068/p3012 Fu, S., Feng, C., Guo, S., Luo, Y., & Parasuraman, R. (2012). Neural adaptation provides evidence for categorical differences in processing of faces and Chinese characters: An ERP study of the N170. PLOS ONE, 7(7), e41103. doi: 10.1371/journal.pone.0041103 Halit, H., de Haan, M., & Johnson, M. H. (2000). Modulation of eventrelated potentials by prototypical and atypical faces. NeuroReport, 11, 1871–1875. doi: 10.1097/00001756-200006260-00014 Han, S. H., Liu, W. Z., Yund, E. W., & Woods, D. L. (2000). Interactions between spatial attention and global/local feature selection: An ERP study. NeuroReport, 11, 2753–2758. doi: 10.1097/00001756-20000 8210-00029 Herrmann, M. J., Ehlis, A. C., Muehlberger, A., & Fallgatter, A. J. (2005). Source localization of early stages of face processing. Brain Topography, 18, 77–85. doi: 10.1007/s10548-005-0277-7 Itier, R. J., Latinus, M., & Taylor, M. J. (2006). Face, eye and object early processing: What is the face specificity? NeuroImage, 29, 667–676. doi: 10.1016/j.neuroimage.2005.07.041 1. F1/F2, FC1/FC2, and C1/C2 were selected for statistical analysis of the N2 component (200–310 ms); F1/Fz/F2, FC1/FCz/FC2, and C1/Cz/C2 were selected for statistical analysis of the P3a component (250–400 ms).
Itier, R. J., & Taylor, M. J. (2002). Inversion and contrast polarity reversal affect both encoding and recognition processes of unfamiliar faces: A repetition study using ERPs. NeuroImage, 15, 353–372. doi: 10.1006/ nimg.2001.0982 Itier, R. J., & Taylor, M. J. (2004a). Effects of repetition learning on upright, inverted and contrast-reversed face processing using ERPs. NeuroImage, 21, 1518–1532. doi: 10.1016/j.neuroimage.2003.12.016 Itier, R. J., & Taylor, M. J. (2004b). N170 or N1? Spatiotemporal differences between object and face processing using ERPs. Cerebral Cortex, 14, 132–142. doi: 10.1093/cercor/bhg111 Joyce, C., & Rossion, B. (2005). The face-sensitive N170 and VPP components manifest the same brain processes: The effect of reference electrode site. Clinical Neurophysiology, 116, 2613–2631. doi: 10.1016/ j.clinph.2005.07.005 Latinus, M., & Taylor, M. J. (2006). Face processing stages: Impact of difficulty and the separation of effects. Brain Research, 1123, 179–187. doi: 10.1016/j.brainres.2006.09.031 Le Grand, R., Cooper, P. A., Mondloch, C. J., Lewis, T. L., Sagiv, N., de Gelder, B., & Maurer, D. (2006). What aspects of face processing are impaired in developmental prosopagnosia? Brain and Cognition, 61, 139–158. doi: 10.1016/j.bandc.2005.11.005 Leder, H., & Carbon, C. C. (2006). Face-specific configural processing of relational information. British Journal of Psychology, 97, 19–29. doi: 10.1348/000712605X54794 Luck, S. J., (2005). The design and interpretation of ERP experiments. In M. S. Gazzaniga (Ed), An introduction to the event-related potential technique (pp. 51–98). Cambridge, MA: MIT Press. Maurer, D., Le Grand, R., & Mondloch, C. J. (2002). The many faces of configural processing. Trends in Cognitive Sciences, 6, 255–260. doi: 10.1016/S1364-6613(02)01903-4 Maurer, D., O’Craven, K. M., Le Grand, R., Mondloch, C. J., Springer, M. V., Lewis, T. L., & Grady, C. L. (2007). Neural correlates of processing facial identity based on features versus their spacing. Neuropsychologia, 45, 1438–1451. doi: 10.1016/j.neuropsychologia. 2006.11.016 Mercure, E., Dick, F., & Johnson, M. H. (2008). Featural and configural face processing differentially modulate ERP components. Brain Research, 1239, 162–170. doi: 10.1016/j.brainres.2008.07.098 Mishra, J., Zinni, M., Bavelier, D., & Hillyard, S. A. (2011). Neural basis of superior performance of action videogame players in an attentiondemanding task. Journal of Neuroscience, 31, 992–998. doi: 10.1523/ JNEUROSCI.4834-10.2011 Morgan, S. T., Hansen, J. C., & Hillyard, S. A. (1996). Selective attention to stimulus location modulates the steady-state visual evoked potential. Proceedings of the National Academy of Sciences, USA, 93, 4770– 4774. doi: 10.1073/pnas.93.10.4770 M€ uller, M. M., & Hillyard, S. (2000). Concurrent recording of steady-state and transient event-related potentials as indices of visual-spatial selective attention. Clinical Neurophysiology, 111, 1544–1552. doi: 10.1016/S1388-2457(00)00371-0
Double dissociation of configural and featural face processing Polich, J. (2007). Updating P300: An integrative theory of P3a and P3b. Clinical Neurophysiology, 118, 2128–2148. doi: 10.1016/j.clinph. 2007.04.019 Renzi, C., Schiavi, S., Carbon, C., Vecchi, T., Silvanto, J., & Cattaneo, Z. (2013). Processing of featural and configural aspects of faces is lateralized in dorsolateral prefrontal cortex: A TMS study. NeuroImage, 74, 45–51. doi: 10.1016/j.neuroimage.2013.02.015 Rossion, B., & Jacques, C. (2008). Does physical interstimulus variance account for early electrophysiological face sensitive responses in the human brain? Ten lessons on the N170. NeuroImage, 39, 1959–1979. doi: 10.1016/j.neuroimage.2007.10.011 Rotshtein, P., Geng, J. J., Driver, J., & Dolan, R. J. (2007). Role of features and second-order spatial relations in face discrimination, face recognition, and individual face skills: Behavioral and functional magnetic resonance imaging data. Journal of Cognitive Neuroscience, 19, 1435– 1452. doi: 10.1162/jocn.2007.19.9.1435 Scott, L. S., & Nelson, C. A. (2006). Featural and configural face processing in adults and infants: A behavioral and electrophysiological investigation. Perception, 35, 1107–1128. doi: 10.1068/p5493 Sreenivasan, K. K., Goldstein, J. M., Lustig, A. G., Rivas, L. R., & Jha, A. P. (2009). Attention to faces modulates early face processing during
1173 low but not high face discriminability. Attention Perception & Psychophysics, 71, 837–846. doi: 10.3758/APP.71.4.837 Ungerleider, L. G., Courtney, S. M., & Haxby, J. V. (1998). A neural system for human visual working memory. Proceedings of the National Academy of Sciences, USA, 95, 883–890. doi: 10.1073/ pnas.95.3.883 Van Essen, D. C., & Gallant, J. L. (1994). Neural mechanisms of form and motion processing in the primate visual system. Neuron, 13, 1–10. doi: 10.1016/0896-6273(94)90455-3 Wang, H., Sun, P., Ip, C., Zhao, X., & Fu, S. (2015). Configural and featural face processing are differently modulated by attentional resources at early stages: An event-related potential study with rapid serial visual presentation. Brain Research, 1602, 75–84. doi: 10.1016/j.brainres. 2015.01.017 Zanto, T. P., & Gazzaley, A. (2009). Neural suppression of irrelevant information underlies optimal working memory performance. Journal of Neuroscience, 29, 3059–3066. doi: 10.1523/JNEUROSCI.4621-08. 2009 (RECEIVED October 30, 2015; ACCEPTED April 6, 2016)
