J Autism Dev Disord DOI 10.1007/s10803-010-1136-2

BRIEF REPORT

Brief Report: Broader Autism Phenotype Predicts Spontaneous Reciprocity of Direct Gaze Frances S. Chen • Jennifer M. D. Yoon

! Springer Science+Business Media, LLC 2010

Abstract We report evidence for a relationship in the general population between self-reported autism-associated traits and the spontaneous reciprocation of direct gaze, a behavior that we propose may reflect a tendency to synchronize with social partners. Adults viewed videos of actors whose gaze was either directed towards or averted from them. Individuals with lower scores on four subscales of the Autism-Spectrum Quotient (AQ) scale showed a greater tendency to look at directed relative to averted eyes; individuals with higher scores on the AQ did not. This relationship was specific to autism-associated traits and to gaze towards the eyes; it did not generalize to a social anxiety measure or to gaze towards the mouth. We discuss implications for our understanding of the broader autism phenotype. Keywords Eye contact ! Direct gaze ! Broader autism phenotype ! Autism-Spectrum Quotient Introduction Eye contact is central to many domains of social cognition, including the recognition of others’ identities, intentions, and emotions; regulation of conversational turn-taking; and judgments about a social partner’s likeability, competence, and credibility (see Itier and Batty 2009; Kleinke 1986). F. S. Chen ! J. M. D. Yoon Department of Psychology, Stanford University, Stanford, California, USA F. S. Chen (&) Department of Psychology, University of Freiburg, Stefan-Meier-Str. 8, 79104 Freiburg, Germany e-mail: [email protected]

Impaired eye contact is a key characteristic of autism (APA 1994) reported in numerous clinical settings and experiments using a range of methodologies (see Senju and Johnson 2009). Atypical eye contact in autism is widely thought to contribute to the pervasive socio-communicative deficits that characterize the disorder, such as impaired recognition of emotions and intentions (Itier and Batty 2009). Although autism has long been associated with atypical gaze behavior, autism has not been consistently associated with an overall reduction in spontaneous eye contact (see Senju and Johnson 2009). Based on the evidence of multiple published observational studies, Buitelaar (1995) concluded that autism is instead characterized by atypical timing of gaze—e.g., a diminished tendency to return a social partner’s direct look into one’s eyes. A possible basis for such reduced reciprocity of direct gaze is suggested by recent behavioral and neurological studies in which individuals with autism, relative to controls, showed atypical detection and processing of direct compared to averted gaze (see Senju and Johnson 2009). Applying eyetracking technology in a controlled laboratory setting to the phenomenon of direct gaze reciprocity promises to provide new evidence about the precise relationship between autism and gaze behavior. Direct gaze reciprocity is also of theoretical interest as a potential example of non-conscious mimicry, which has been shown to influence the quality of social relationships (Chartrand and Bargh 1999). It may also serve as an indicator of an individual’s general attentiveness to, and tendency to synchronize with, changes in the social environment. Furthermore, direct gaze reciprocity may be a teachable behavior, potentially facilitating social interactions and improving emotion-reading and interpretation of mental states.

123

J Autism Dev Disord

The general population is an accessible starting point for investigating the relationship between direct gaze reciprocity and autism-associated traits. A broader autism phenotype (BAP), which includes milder manifestations of the socio-communicative difficulties and preferences for restricted activities that characterize autism, exists in the general population (Baron-Cohenet al. 2001). The BAP has been associated with high neuroticism, low extraversion, depression, and anxiety (Austin 2005; Kunihira et al. 2006), as well as impaired interpretation of body language and facial expressions (Ingersoll 2010). While the aforementioned evidence links the BAP to self-reported activities, personality traits, and cognitive impairments, evidence about actual behaviors that characterize the BAP is relatively lacking. Evidence that gaze behavior and autism-associated traits are correlated in the general population would provide a novel link between the BAP and an observable behavioral tendency. Although siblings of individuals with autism are known to show atypical eye contact early in childhood (Dalton et al. 2007), it remains unknown whether atypical eye contact and the BAP are connected in non-relatives of individuals with autism. In our study, we recruited adults from the general population and recorded their gaze behavior as they viewed videos of actors whose gaze was either directed towards or averted from them. We predicted that high levels of autism-associated characteristics would correlate with a decreased tendency to reciprocate the direct gaze of these videotaped social partners.

directly into the video camera) and an averted condition (head and eye gaze averted by 45 degrees). Fifty-two undergraduates viewed one direct-gaze and one averted-gaze video (order counterbalanced) of two different actors matched to their own sex. Participants were instructed to watch naturally. An ASL6000 eyetracker collected eye position information. Participants also completed the Social Skill, Communication, Attention to Detail, and Attention Switching subscales of the AutismSpectrum Quotient (AQ), a questionnaire that assesses features of the BAP in healthy adults (Baron-Cohen et al. 2001). Six students who recognized one of the actors, and eight whose eyetracking data were missing or miscalibrated, were excluded, leaving 38 subjects for analysis. We analyzed percentage of total time spent in fixations within the eye region, defining a fixation when the standard deviation of the x and y gaze coordinates was below 0.5" visual angle for at least 100 ms. Modulation of gaze behavior was operationalized as the difference between subjects’ scores in the direct and averted conditions. We analyzed our data in two ways: dividing subjects into low and high AQ groups based on a median split of questionnaire scores (excluding three median-equivalent subjects), and conducting correlation analyses including all 38 subjects. We included two control measures—gaze to the mouth, and a measure of social anxiety (the BFNE, Leary 1983)— to investigate the specificity of the hypothesized relationship between direct gaze reciprocity and autism-associated characteristics.

Methods

Results

We created eight naturalistic two-minute videos of four undergraduates (two male, two female) speaking about neutral topics such as their daily routines. Each actor spoke about the same topic under a direct condition (looking

Demographic data are shown in Table 1. Preliminary analyses revealed no significant effects of participant sex on AQ scores or gaze behavior, and the variable was excluded from further analyses.

Table 1 Participant characteristics Overall N

38

Age in years Sex (M:F)

18:20

Abbreviated AQ score

Low AQ group

High AQ group

Group comparison

16

19

19.1 (±0.8)

19.1 (±1.0)

t(33) = 0.6, ns

6:10

10:9

v2 = 0.8, ns

14.3 (±5.7)

9.1 (±2.2)

18.8 (±4.1)

t(33) = 8.6*

Social Skills subscale

2.4 (±2.0)

1.1 (±1.3)

3.8 (±1.7)

t(33) = 5.2*

Communication subscale

2.2 (±2.0)

0.6 (±0.7)

3.6 (±1.6)

t(33) = 6.8*

Attention Switching subscale

4.8 (±2.0)

3.5 (±1.4)

5.9 (±2.0)

t(33) = 4.1*

Attention to Detail subscale

4.8 (±1.7)

4.0 (±1.5)

5.5 (±1.5)

t(33) = 3.0*

The ‘‘Overall’’ group includes three participants with median-equivalent scores (=13) not included in either the low or high AQ group * p \ .01

123

J Autism Dev Disord

The percentages of time that each group spent looking at direct and averted eyes are shown in Fig. 1, Panel A. We conducted a repeated-measures ANOVA with group (low versus high AQ) as a between-subjects factor, condition (direct versus averted video) as a repeated factor, and viewing order as a covariate. Only the group-by-condition interaction was significant (F(1, 32) = 4.60, p \ .05, g2 = .13). Planned within-group contrasts revealed that individuals who reported relatively few autism-associated characteristics (low AQ) spent a greater percentage of time looking at direct than averted eyes (mean difference = 9.1%, SD = 17.1%, t(15) = 2.14, p \ .05, d = .53), whereas individuals who reported relatively more autism-associated characteristics (high AQ) did not (mean difference = -3.6%, SD = 16.7%, t(18) = .94, ns). The correlation between AQ score and modulation of percentage time on eyes was also significant in the predicted direction (r = .28, p \ .05, one-tailed). Thus, individuals reporting fewer autism-associated traits showed gaze modulation as a function of the actor’s gaze direction, while individuals reporting more autism-associated traits did not. To further specify how gaze behavior of high AQ and low AQ participants differed, we also analyzed two independent elements of our overall gaze measure: (1) length of looks, i.e., mean duration of individual fixation events within the eye region, and (2) frequency of looks, i.e., percentage of fixations within the eye region (relative to total fixations). The first measure may be interpreted as a tendency to linger on, or be captured by, a stimulus, whereas the second measure may be interpreted as a tendency to orient repeatedly towards a stimulus.

Mean durations of fixations toward direct and averted eyes are shown in Fig. 1, Panel B. As before, we conducted an ANOVA with group as a between-subjects factor, condition as a repeated factor, and viewing order as a covariate. There were no significant main effects, but the group-bycondition interaction was significant (F(1, 32) = 5.64, p\.05, g2 = .15). Planned within-group contrasts revealed that low AQ individuals made relatively longer individual fixations to direct versus averted eyes (mean difference = 225 ms, SD = 324 ms, t(15) = 2.77, p \ .05, d = .69), whereas high AQ individuals did not (mean difference = 5 ms, SD = 238 ms, t = .10, ns). The correlation between AQ score and modulation of fixation length to eyes was also significant in the predicted direction (r = .41, p\.01, one-tailed). There was an effect of presentation order (F(1, 32) = 4.66, p \ .05, g2 = .13) such that subjects viewing the direct video first made longer individual fixations to the eyes throughout the experiment. This result suggests that viewing direct gaze may affect an individual’s subsequent attention and gaze behavior, a potentially interesting avenue for exploration in future studies. In the current study, however, this effect did not interact with our primary variable of interest—group differences—nor was it significant in the previous analysis of overall gaze behavior. The groups did not differ in the frequency with which they looked towards direct versus averted eyes. Our control analyses revealed no correlations between gaze modulation to the mouth and AQ scores, nor was social anxiety correlated to any measure of gaze modulation.

Discussion

Fig. 1 Gaze patterns toward direct and averted eyes. Error bars represent one within-subject standard error

We found evidence for a relationship between direct gaze reciprocity and self-reported autism-associated traits in the general population. Individuals with lower scores on four subscales of the Autism-Spectrum Quotient showed a greater spontaneous tendency to look at directed relative to averted eyes. This gaze modulation effect appears to be selectively linked to the broader autism phenotype (BAP), as it did not generalize to a measure of social anxiety. To our knowledge, this result represents the first documented link between the BAP and this systematic individual difference in gaze behavior. Analyses of two separate components of gaze modulation—length of individual looks and frequency of looks— suggest that low AQ individuals modulate their gaze primarily by making longer individual fixations to direct versus averted eyes. Further research should investigate these distinct elements of gaze behavior and the different mechanisms that may control them. Direct gaze reciprocity is theoretically relevant for autism research not only as a type of gaze behavior, but also

123

J Autism Dev Disord

because it includes an element of imitation or non-conscious mimicry. Individuals with Autism Spectrum Disorder show limited behavioral imitation (Williams et al. 2004) and limited automatic mimicry of emotional facial expressions (McIntosh et al. 2006). Further research might examine whether individuals who show reduced direct gaze reciprocity also show reduced mimicry in other contexts, and whether direct gaze reciprocity correlates to an individual’s attention to other changes in the social environment. Furthermore, an intervention based on teaching direct gaze reciprocity could reveal whether a causal relationship exists between direct gaze reciprocity and the socio-communicative deficits associated with autism. We did not find a correlation between autism-associated characteristics and absolute levels of looking at either direct or averted eyes. Instead, our results suggest that the BAP is associated with the more specific behavioral tendency to modulate one’s own gaze as a function of a social partner’s gaze direction. Future research with a clinicallydiagnosed sample is needed to determine whether individuals with autism also demonstrate reduced gaze modulation under these conditions, and to extend the results beyond those currently obtained with a self-report measure of autism-associated traits. We measured autism-associated traits using only four subscales of the Autism Spectrum Quotient. This leaves open whether direct gaze reciprocity is linked to other aspects of the BAP, including the ‘‘Imagination’’ subscale that we omitted. Future studies in this line of research should use the full Autism Spectrum Quotient to link direct gaze reciprocity more directly to other studies using the AQ. Furthermore, the inclusion of a broader range of visual stimuli (such as opposite-sex actors, familiar individuals, or individuals displaying various emotional expressions) will help to determine the generalizability of the effect we found. Various research methodologies have shown significant overlap in how people respond to live versus filmed or digital representations of social partners (cf. Reeves and Nass 1996). Our methodology, which uses a live participant and a videotaped actor, has the important strength that superficial features of the stimulus, such as tone of voice and speech content, are held constant. By controlling for these features our study represents a necessary initial step towards the broader goal of understanding live interactions. In live interactions, gaze functions as part of a two-way communicative system, as both a response to, and signal for, the social partner. Moment-to-moment changes in a social agent’s gaze behavior are likely to have important consequences on the social partner’s perceptions of the agent’s likeability, attentiveness, trustworthiness, and social skills. We hope that this study sparks further research elucidating precise ways in which direct gaze

123

reciprocity influences the intricate feedback cycle of live social interaction. Acknowledgments The authors wish to thank Susan Johnson for helpful guidance on our experimental design and data analysis, and Stefan Sperling for generous feedback on previous versions of the manuscript. The authors also gratefully acknowledge the technical assistance of Lucas Manfield, Heidi Fogle, Antonia Garcia, Joey Klein, and Katie O’Neil.

References American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: American Psychiatric Association. Austin, E. J. (2005). Personality correlates of the broader autism phenotype as assessed by the Autism Spectrum Quotient (AQ). Personality and Individual Differences, 38(2), 451–460. Baron-Cohen, S., Wheelwright, S., Skinner, R., Martin, J., & Clubley, E. (2001). The Autism-spectrum Quotient (AQ): Evidence from Asperger syndrome/high functioning autism, males and females, scientists and mathematicians. Journal of Autism and Developmental Disorders, 31(1), 5–17. Buitelaar, J. K. (1995). Attachment and social withdrawal in autism: hypotheses and findings. Behaviour, 132, 319–350. Chartrand, T. L., & Bargh, J. A. (1999). The chameleon effect: The perception-behavior link and social interaction. Journal of Personality and Social Psychology, 76(6), 893–910. Dalton, K., Nacewicz, B., Alexander, A., & Davidson, R. (2007). Gaze-fixation, brain activation, and amygdala volume in unaffected siblings of individuals with autism. Biological Psychiatry, 61(4), 512–520. Ingersoll, B. (2010). Broader autism phenotype and nonverbal sensitivity: Evidence for an association in the general population. Journal of Autism and Developmental Disorders, 40(5), 590–598. Itier, R. J., & Batty, M. (2009). Neural bases of eye and gaze processing: The core of social cognition. Neuroscience and Biobehavioral Reviews, 33, 843–863. Kleinke, C. L. (1986). Gaze and eye contact: a research review. Psychological Bulletin, 100(1), 78–100. Kunihira, Y., Senju, A., Dairoku, H., Wakabayashi, A., & Hasegawa, T. (2006). ‘Autistic’ traits in non-autistic Japanese populations: Relationships with personality traits and cognitive ability. Journal of Autism and Developmental Disorders, 36(4), 553–566. Leary, M. R. (1983). A brief version of the Fear of Negative Evaluation scale. Personality and Social Psychology Bulletin, 9(3), 371–375. McIntosh, D. N., Reichmann-Decker, A., Winkielman, P., & Wilbarger, J. L. (2006). When the social mirror breaks: deficits in automatic, but not voluntary, mimicry of emotional facial expressions in autism. Developmental Science, 9, 295–302. Reeves, B., & Nass, C. (1996). The media equation: How people treat computers, television, and new media like real people and places. Stanford, CA: CSLI Publications/Cambridge University Press. Senju, A., & Johnson, M. H. (2009). Atypical eye contact in autism: Models, mechanisms and development. Neuroscience and Biobehavioral Reviews, 33, 1204–1214. Williams, J. H. G., Whiten, A., & Singh, T. (2004). A systematic review of action imitation in Autistic Spectrum Disorder. Journal of Autism and Developmental Disorders, 34, 285–299.