What is social perception and attribution?

Social Perception versus Faces

The Social Perception and Faces subtests both require processing facial information. When patients have low scores on Social Perception and Faces, they may have a general deficit in face processing that may be affecting their performance on the Social Perception test. In contrast, patients whose Social Perception scores are low and whose Faces scores are within normal limits are likely to have intact face processing, suggesting a specific deficit in emotion recognition or processing of emotional content. Lastly, low scores on Faces with Social Perception scores within normal limits suggest difficulties with face memory, but not necessarily with other aspects of face discrimination.

Social Perception

Social Perception measures comprehension of social communication, including facial affect recognition and naming, affect recognition from prosody and facial expressions, and affect recognition from prosody and interaction between people. Three tasks comprise Social Perception: Affect Naming, Prosody-Face Matching, and Prosody-Pair Matching. In Affect Naming, the examinee is shown photographs of faces and selects an emotion from a card to describe the affect demonstrated in the photograph. In Prosody-Face Matching, the examinee hears an audio-recorded statement and selects one face from four choices that matches the emotion expressed in the recording. In Prosody-Pair Matching, the examinee hears an audio-recorded statement and selects one photograph of interacting pairs of individuals from four choices that matches the meaning of the speaker’s statement. For Prosody-Face Matching and Prosody-Pair Matching, the statement content may not match the emotion expressed. This intentional lack of matching allows for better measurement of more subtle forms of communication, such as sarcasm. Available scores assess the examinee’s overall performance, as well as the individual skills measured within and across the three tasks.

Social perception

Social perception is typically partitioned into two categories, namely facial perception and vocal perception (Green et al., 2015). Vocal perception involves recognizing and discriminating acoustic properties of speech, and the affective information they convey, where facial perception is decoding affective information from others’ facial expressions. Inherent in both categories is the ability to read others’ emotional cues. Generally speaking, there is evidence that individuals with schizotypy show mildly impaired performance on tasks of social perception, though the nature of these impairments has not replicated well across studies. For example, van’t Wout, Aleman, Kessels, Larøi, and Kahn (2004) administered both the SPQ and a facial affect recognition task, and found that positive schizotypy correlated with misclassifying angry faces as happy, and the subscale of unusual perceptual experiences correlated with misclassifying happy faces as angry or fearful. Other studies have extended these findings and observed that negative schizotypy is also related to deficits in facial affect (e.g., Abbott & Green, 2013; Williams, Henry, & Green, 2007). However, some studies do not replicate deficits in classifying facial emotion (e.g., Jahshan & Sergi, 2007; Shean, Bell, & Cameron, 2007), but these null findings are potentially limited by the emotion perception tasks used, which may be insufficiently sensitive. Brown and Cohen (2010) sought to use a more sensitive measure of facial emotion perception that included neutral stimuli, and found that people with schizotypy were worse at identifying emotions than controls, more likely to misclassify neutral faces as showing disgust, and that there was a relation between disorganized schizotypy and a systematic bias for classifying faces with a negative emotional valence. Individuals with schizotypy also show deficits in facial recognition, as demonstrated by Larøi, D’Argembeau, Brédart, and van der Linden (2007), who found that people higher in schizotypy traits endorsed greater difficulty recognizing themselves, and others they know, and also show perceptual distortion of faces, in comparison with people low in schizotypy. These difficulties and distortions were associated with positive and disorganized, but not negative, schizotypy. In sum, evidence suggests that there is a facial affect recognition negativity bias in people with schizotypy, and that kinds of schizotypy may differentially misread various emotions, which may be driven by a lack of ability to integrate facial cues more broadly.

These results are replicated in studies on emotion recognition from vocal information. Although there are significantly fewer studies overall, they consistently show deficits in vocal affect recognition, the ability to recognize emotions from speech, in schizotypy samples (Castro & Pearson, 2011), as well as in individuals diagnosed with schizotypal personality disorder (Baum & Nowicki, 1998; Wickline, Nowicki, Bollini, & Walker, 2012), and at ultra-high risk for psychosis (Amminger et al., 2012). One study partially extended this finding, with the specification that positive (particularly cognitive-perceptual) schizotypy was negatively related to vocal emotion recognition accuracy (Shean et al., 2007). Taken together, these studies and the studies of facial affect recognition show a deficit in social perception in schizotypy that mirrors the results of similar studies conducted in schizophrenia (Kohler et al., 2003; Tseng et al., 2013). Some researchers have reported a negative correlation between ability to identify happy emotions and both negative and positive symptoms (Tseng et al., 2013). In contrast, Kohler et al. (2003) observed that facial affect recognition in schizophrenia was worse relative to controls for fearful, disgusted, and neutral faces, but not for happy or angry faces, and worse performance was connected specifically to negative symptoms. Similar to facial emotion recognition, there is a pattern of general deficit in social perception, but mixed results at the level of what kinds of perceptions are most difficult, and which symptom profiles are most closely connected.

6 Conclusion

Our knowledge of social perception was once limited to its downstream consequences, such as how categorization and perception influence stereotyping and behavior.40,152 The past decade has seen unprecedented progress in unveiling the processes underlying initial percepts. Such progress has been driven by an equally unprecedented integration between disciplines, including the anticipated wedding of social-cognitive and neural sciences,162–164 as well as social-cognitive and visual sciences.1,2,165 Together, these perspectives have engendered productive theoretical accounts of social perception36 and have galvanized research into its basis across levels of analysis. In this chapter, we have focused on the top-down influences in social perception, and importantly, the interesting and nuanced ways in which they interact with different levels of processing. Specifically, we have reviewed neuroscience research into top-down influences in social vision, a parsimonious account of these influences from the vision and neural sciences, and have discussed how these areas may inform one another and fit within current computational frameworks. We hope the current direction of this research forges a productive collaboration that informs each of the social, cognitive, and neural sciences.

Social Perception

Social perception refers to identifying and utilizing social cues to make judgments about social roles, rules, relationships, context, or the characteristics (e.g., trustworthiness) of others. This domain also includes social knowledge, which refers to one’s knowledge of social roles, norms, and schemas surrounding social situations and interactions. Tests of social perception include videotaped scenes that require the viewer to make inferences and judgments about ambiguous social situations based on limited verbal and nonverbal social cues (e.g., Profile of Nonverbal Sensitivity (PONS); Rosenthal et al., 1979; Social Cue Recognition Test (SCRT); Corrigan, 1997). The Relationships across Domains task (RAD; Sergi et al., 2009) requires participants to make inferences about the nature of relationships between people based on short written vignettes.

1.10.2.1 An Essential Base

The topics of social perception and attribution occupy a central role in research and practice in clinical psychology. As will be shown in a later discussion of Rosenhan's famous study “On being sane in insane places” (Rosenhan, 1973), social perception, or how we perceive other people, is central to the understanding of labeling behavior as abnormal. Rosenhan essentially showed that even educated professionals, too quickly and without cautious scrutiny of the evidence, label others and then direct behavior toward them that corresponds with the label. In so doing, he revealed a major feature of social perception, namely that it often is biased in terms of its rational, careful consideration of available information. Beyond Rosenhan's demonstration, it has been recognized since the mid-twentieth century that social perception and attribution, which refers to inferences about causality and responsibility, are pivotal in determining how people interact with one another. These processes, hence, are critical not only in the unfolding of a person's social interactions in all areas of life, but also in therapist–client interactions during therapy.

In social psychology, social perception often is treated as the general area that subsumes a number of sub areas, including attribution, social cognition, impression formation, person perception, and self-perception (Weary, Stanley, & Harvey, 1989). But since the late 1950s, each of these subtopics has spawned a large, specialized literature. It has been a long time since scholars such as Heider (1958) and Kelley (1967) attempted to integrate the diverse ideas and findings. Nevertheless, their earlier analyses still have great reach regarding the development of theory and research pertaining to people's perception of others.

The concept of self is a general topic that traditionally was considered separately from social perception (James, 1890). However, self-perception, which refers to the inferences we make about ourselves, was conceived by Bem (1972) as involving a process in which people learn about themselves and their attitudes and dispositions by engaging in an attribution process not unlike what they do when they make inferences about others. In addition, Jones (1964) built on Goffman's (1959) ideas about self-presentation and developed a set of penetrating analyses of how people tactically present themselves to others in order to try to fulfill various goals of social interaction. For example, people may engage in ingratiation, self-promotion, self-martyring, and so on with the goal of social or material gain.

Early work on social perception demonstrated the remarkable degree to which people's own self-biases and motivations affect their perceptions of others. A classic example of this tendency toward bias in perception was provided by Hastorf and Cantril (1954). They showed students at Dartmouth and Princeton Universities film from a football game played in 1951 between those two teams. The game was rough, with lots of penalties especially against Dartmouth, and Princeton's star All-American selection Dick Kazmaier was injured in the rough play. Princeton won but commentary in the local papers continued for some time about the disgusting display of poor sportsmanship.

Hastorf and Cantril's procedure involved showing students at both schools the same portions of the game film and asking them to indicate on a questionnaire the rough play infractions they saw. Hastorf and Cantril found that Princeton students showed a very strong tendency to judge Dartmouth players as most responsible for the rough play. On the other hand, Dartmouth students de-emphasized their players' rough play and judged the teams as generally equally guilty of infractions. Thus, there is historical evidence for the idea that people can observe the same stimuli and come to different judgments based on their motivations or self-biasing positions. Such evidence is important both to therapists who must make judgments about the dispositions of patients/ clients and also to therapists and clinical scholars interested in analyzing the distortions in people's perceptions of others in their social environment.

2.11.5.2.3 Knowledge statements related to social and multicultural bases of behavior

(i)

Social cognition and perception (e.g., attribution theory and biases, information integration, confirmation bias, person perception, development of stereotypes, racism).

Social interaction (e.g., interpersonal relationships, aggression, altruism, attraction).

Group dynamics and organizational structures (e.g., school systems, gang behavior, family systems, group thinking, cultural behavior, conformity, compliance, obedience, persuasion) and social influences on individual functioning.

Environmental/ecological psychology (e.g., person–environment fit, crowding, pollution, noise).

Theories of personality that describe behavior and the etiology of atypical behavior. Includes knowledge of limitations in existing theories for understanding the effect of diversity (e.g., age, ethnicity, gender).

Multicultural and multiethnic diversity (e.g., racial/ethnic minorities, gender, age, disability, sexual orientation, religious groups, between- and within-group differences).

Theories of identity development of multicultural/multiethnic groups (e.g., acculturation theories, racial/ethnic identity).

Role that race, ethnicity, gender, sexual orientation, disability, and other cultural differences play in the psychosocial, political, and economic development of individuals/groups.

Sexual orientation issues (e.g., sexual identity, gay/lesbian/bisexual, family issues).

Psychology of gender (e.g., psychology of women, psychology of men, gender identity development).

Disability and rehabilitation issues (e.g., inclusion, psychological impact of disability).

Social perception

Substantially less work has explored social perception in FEP, however, findings once again are largely consistent in reporting deficits in FEP groups relative to healthy controls (Healey et al., 2016). However, many studies did not control for the effect of IQ, and given that social perception has been shown to significantly correlate with neurocognition (Addington, Saeedi, & Addington, 2006b), further work is necessary to determine whether deficits in social perception exist independent of general cognitive impairment in FEP.

The few studies that have investigated the stability of social perception deficits in FEP found performance to be stably impaired over a 1-year (Addington et al., 2006b; Horan et al., 2012) and 5-year period (McCleery et al., 2016). In a study by Addington et al. (2006b), the healthy control group showed significant improvement in social perception performance over time, while the FEP group did not. This suggests that the FEP individuals are not exhibiting the same developmental improvements as seen in healthy individuals. Further work mapping the trajectory of social cognition in healthy cohorts is necessary (Healey et al., 2016).

Social Categorization

A core aspect of social perception is thinking about people in terms of their social group memberships – this is called social categorization. Social categorization can occur reflexively within milliseconds of perception to disambiguate a complex social world. Perceiving an individual in terms of their social category membership allows perceivers to apply preexisting information and attitudes about that social category to the individual. By providing information about otherwise novel individuals, social categorization reduces uncertainty about the social world and helps perceivers anticipate the characteristics of others and calibrate their own behavior. Although social categorization is necessary to make informed inferences about novel people, its reliance on cognitive shortcuts such as heuristics, stereotypes, and prejudices may lead to systematic inaccuracies in the processing of any given person. For example, categorizing someone as a fraternity member may, on average, provide useful predictive information about their interests and social behavior, but there is no guarantee that these predictions will apply to this particular member.

A long tradition in intergroup relations has documented the cognitive and affective implications of perceiving in-group versus out-group members. One of the most robust effects of group membership is called the same-race memory bias. The same-race memory bias is a phenomenon whereby individuals are more likely to remember the identity of members of their own racial group relative to the members of other racial groups. The same-race memory bias has been extended to variety of nonracial groups (including novel groups formed under minimal conditions), strongly suggesting that the phenomenon is driven by deeper encoding of individual in-group members, regardless of race. It appears that the fusiform gyrus, an area of the brain involved in facial and categorical processing, may play a key role in the same-race memory bias. When viewing racial in-group members, Black and White individuals have stronger activation in the fusiform face area, an area of the fusiform sensitive to faces that may be more generally engaged in expert, in-depth processing of categories of stimuli. Moreover, activation of the fusiform to racial in-group (>out-group) members is correlated with the degree of same-race memory bias (i.e., superior memory for in-group faces). This research suggests that lower-level perceptual processing may play a key role in explicit racial biases in memory.

To help understand more about unconscious intergroup processing, researchers have started to explore the timecourse of neural responses to social groups. A number of recent studies have used event-related potentials (ERPs) to examine the automatic and controlled components of intergroup social categorization and evaluation. Whereas very early ERP waves (c. 170 ms) distinguish facial from nonface stimuli, subsequent ERP waves respond differentially to in-group and out-group members. Indeed, several studies have shown that racial in-group members are associated with larger ERP waves and thus more processing approximately 250 ms following stimulus presentation. These data converge with research on the same-race memory bias, suggesting that social categorization occurs rapidly and may be mediated by perceptual processes, and perhaps activity in the fusiform gyrus. Indeed, the rapidity with which individuals differentiate in-group from out-group members raises the possibility that early social categorization processes may be occurring outside conscious awareness and control. These data are also consistent with the view that individuals often engage in more in-depth processing of in-group members.

Consistent with these in-group biases in the central nervous system, similar biases are evident in PNS responses. Although activity in the PNS is generally slower than the central nervous system, it can nevertheless provide important insights into aspects of intergroup processing, such as social categorization. Converging with fMRI and EEG studies, PNS activity is heightened while perceiving the emotional displays of in-group members. Specifically, Black and White individuals both show increased skin conductance and facial EMG responses to pleasant and unpleasant pictures of racial in-group (>out-group) members. In summary, these studies provide evidence that perceptual processes may be automatically biased toward in-group members and that these early attentional differences may have implications for downstream memory and PNS activity.

Perceiving Trustworthiness in Faces

Traditionally, neuroimaging studies on social perception have associated the evaluation of facial trustworthiness with the amygdala (Todorov, Mende-Siedlecki, & Dotsch, 2013). Some of the earliest evidence for this link comes from neuropsychological research conducted by Adolphs, Tranel, and Damasio (1998), which demonstrated clear dissociations between patients with bilateral amygdala damage and normal controls with respect to evaluations of facial trustworthiness. Specifically, the bilateral amygdala patients rated untrustworthy-looking faces as being more trustworthy and approachable than did controls, brain-damaged controls, or unilateral amygdala lesion patients. These authors suggest that this difference can be attributed to deficits in extracting relevant social information from the face, as opposed to a more general decrement in face processing (Adolphs et al., 1998). Indeed, these results can be contrasted with more recent work suggesting that both acquired and developmental prosopagnosics provide standard trustworthiness evaluations (Quadflieg, Todorov, Laguesse, & Rossion, 2012; Todorov & Duchaine, 2008).

In the intervening years, numerous fMRI investigations have corroborated these initial findings, as have subsequent meta-analyses (Bzdok et al., 2011; Mende-Siedlecki, Said, & Todorov, 2013), which affirm the amygdala's contributions to the social evaluation of faces across numerous behavioral tasks and stimuli sets. That being said, this additional work has added new wrinkles to the narrative. For example, while the earliest of these neuroimaging studies observed a negative relationship between facial trustworthiness and amygdala activity (Engell, Haxby, & Todorov, 2007; Winston, Strange, O'Doherty, & Dolan, 2002), such that the amygdala responded most strongly to untrustworthy-looking faces, more recent evidence points to a nonlinear relationship between amygdala activity and facial trustworthiness. Across a growing number of studies, researchers have observed enhanced amygdala activity to both untrustworthy- and trustworthy-looking faces, as compared to the more neutral faces towards the midpoint of the trustworthiness dimension (Mattavelli, Andrews, Asghar, Towler, & Young, 2012; Said, Baron, & Todorov, 2009; Said, Dotsch, & Todorov, 2010; Todorov, Baron, & Oosterhof, 2008; Todorov, Said, Oosterhof, & Engell, 2011).

A parallel line of research has sought to address the temporal dynamics of facial trustworthiness evaluation by measuring event-related potentials (ERPs; Dzhelyova, Perrett, & Jentzsch, 2012; Kovács-Bálint, Stefanics, Trunk, & Hernádi, 2014; Marzi, Righi, Ottonello, Cincotta, & Viggiano, 2014; Rudoy & Paller, 2009; Yang, Qi, Ding, & Song, 2011). Though the initial findings of these studies are somewhat mixed, taken together, they suggest that facial trustworthiness exerts an early influence on visual processing (Dzhelyova et al., 2012; Marzi et al., 2014), dovetailing with behavioral work suggesting that evaluations of facial trustworthiness are made after only 100 milliseconds of exposure to faces (Willis & Todorov, 2006). Moreover, these studies demonstrate enhanced processing associated with untrustworthy faces, which persists throughout later latencies, potentially reflecting subsequent motivational consequences triggered by trustworthiness evaluation (Marzi et al., 2014; Yang et al., 2011). Finally, these authors posit a similarity between ERP responses to facial trustworthiness and previously observed responses associated with processing of facial expressions (Dzhelyova et al., 2012; Marzi et al., 2014; Yang et al., 2011). This finding has been interpreted as further evidence for the emotion overgeneralization hypothesis (Marzi et al., 2014; Yang et al., 2011), which suggests that the social evaluation of faces has emerged as an extension of the processes supporting the recognition and comprehension of emotional facial expressions (Montepare & Dobish, 2003; Oosterhof & Todorov, 2008; Said, Sebe, & Todorov, 2009; Todorov, 2008; Zebrowitz & Montepare, 2008). In fact, computational modeling of face trustworthiness shows that whereas untrustworthy-looking faces resemble angry faces, trustworthy-looking faces resemble happy faces (Oosterhof & Todorov, 2008).

Finally, we note that incidental cues associated with faces can influence evaluations of facial trustworthiness, as well. For example, peripheral distractor faces in a visual search paradigm (Frischen, Ferrey, Burt, Pistchik, & Fenske, 2012; Raymond, Fenske, & Westoby, 2005), faces associated with unreliable eye gaze in a gaze-cueing task (Bayliss & Tipper, 2006), and faces associated with response inhibition in a go/no-go task (Fenske, Raymond, Kessler, Westoby, & Tipper, 2005) are all rated as looking less trustworthy than corresponding control faces. The neural bases of the latter effect were recently uncovered by Doallo et al. (2012), who observed that not only were no-go faces perceived to be less trustworthy but also they elicited enhanced activity in the amygdala during a subsequent recognition task. Moreover, both behavioral and neural indices of this devaluation were predicted by lateral orbitofrontal cortex (OFC) activity, such that greater recruitment of this inhibitory control mechanism during the go/no-go task was associated with both less trustworthy evaluations of and greater amygdala activity in response to no-go faces.