Eye contact and arousal in WS   Stuck on you: Face-to-face arousal and gaze aversion in Williams syndrome Gwyneth Doherty-Sneddon,1 Deborah M Riby,2 Lesley Calderwood,3 Leanne Ainsworth1 1Dept. of Psychology Stirling University 2School of Psychology, Newcastle University 3Dept. of Psychology, Glasgow Caledonian University Keywords: Williams syndrome, eye contact, gaze, arousal Correspondence should be addressed to: Dr. Gwyneth Doherty-Sneddon, Department of Psychology, University of Stirling, FK9 4LA, U.K. Email [email protected] Tel: + 44 (0) 1786 467653 Fax: + 44 (0) 1786 467641 1 Eye contact and arousal in WS   Abstract Introduction: During face-to-face questioning typically developing children and adults use gaze aversion (GA), away from their questioner, when thinking. GA increases with question difficulty and improves the accuracy of responses. We investigate whether individuals with Williams syndrome (WS), associated with hyper-sociability and atypical face gaze, use GA to manage cognitive load and whether physiological arousal is associated with looking at faces. Methods: Two studies were conducted by: i) Recording changes in the participants’ skin conductance levels whilst manipulating task difficulty and gaze direction and ii) Calculating the amount of GA away from the experimenters’ face whilst answering questions of varying difficulty. Results: In study 1, whilst WS was associated with general hypo-arousal, face arousal effects were found for both Williams syndrome and typically developing participants. In study 2, participants with WS showed prolonged face gaze under high task demands, however question difficulty did increase GA. Conclusions: Looking at faces is demanding, even for individuals with WS. Decreased physiological arousal may allow individuals with WS hold face gaze for prolonged periods of time, but looking at faces does increase baseline arousal level. The results are discussed in terms of social skills training and teaching methods appropriate for WS. 2 Eye contact and arousal in WS   Stuck on you: Face-to-face arousal and gaze aversion in Williams syndrome Williams syndrome (WS) is a genetic disorder arising from a microdeletion of approximately 25 genes at chromosome site 7q11.23 (Donnai & Karmiloff-Smith, 2000). Previous research has mapped the distinct neuropsychological profile associated with WS, showing strengths and weaknesses of cognitive ability across verbal and nonverbal domains of processing (Bellugi, Wang & Jernigan, 1994, Bellugi, Lichtenberger, Jones, Lai, & St George, 2000). Interest in the social phenotype associated with WS has increased exponentially over recent years with suggestions of subtle atypicalities, defined as ‘hyper-sociability’ (Jones et al., 2000, Doyle, Bellugi, Korenberg, & Graham, 2004) or a ‘pro-social drive’ (Frigerio et al., 2006). Individuals with WS are often characterised as overfriendly and unreserved with both familiar and unfamiliar people (Gosch & Pankau, 1997). Children with WS rate unfamiliar faces as more approachable than children with other disorders of development (Jones et al., 2000), especially when they show a happy expression (Frigerio et al., 2006) and when engaged in social interactions toddlers with WS show intense looking behaviour towards faces (Mervis et al.., 2003). Due to the social relevance of interpreting communicative facial cues and the interest that individuals with WS show in looking at faces, behavioural studies have explored face perception with the aim of understanding interpersonal communication. Research suggests that although individuals may appear relatively proficient at recognising faces (Bellugi et al., 2000) they exhibit a range of atypicalities regarding structural encoding (Karmiloff-Smith et al., 2004) and specifically the encoding of familiar faces (Riby, Doherty-Sneddon, & Bruce, 2008a). Within a profile of skills across disorders of development, individuals with WS show 3 Eye contact and arousal in WS   significantly more competence at interpreting cues of eye gaze and expressions of emotion when compared to individuals with autism (Riby, Doherty-Sneddon, & Bruce, 2008b). The difference in interpreting communicative facial cues is likely to contribute to the socio- cognitive differences shown by individuals with WS and those with autism (Brock, Einav & Riby, 2008). Indeed individuals with WS appear relatively proficient at reading simple and complex mental states from the eye region (Tager-Flusberg, Boshart, & Baron-Cohen, 1998), a problem that is difficult for those with autism (Baron-Cohen, Leslie, & Frith, 1986). Recent eye-tracking investigations have shown that when individuals with WS look at static pictures of people in social situations (e.g. pictures showing characters chatting to their friends, or making lunch) they spend longer looking at actors’ faces than participants who are developing typically (Riby & Hancock, 2008). Increased fixation duration is particularly apparent towards the eye region of characters, linking to claims of prolonged face gaze and eye contact (Mervis et al., 2003). This finding is replicated when individuals with WS watch movies containing humans, but not when watching cartoons (Riby & Hancock, 2009). Eye- tracking research therefore corroborates suggestions that the gaze behaviour of individuals with WS is atypical, evidenced by increased and prolonged face and eye fixations towards human faces. Current explanations for the occurrence of hyper-sociable behaviour (and exaggerated face gaze) focus on a number of possible contributors. The role played by the amygdala (e.g. Jawaid, Schmolck & Schulz, 2008) has attracted research attention due to abnormalities of structure and size (Reiss et al., 2004) in individuals with WS. The important role of the amygdala in processing information related to emotion and its link to socio- emotional functioning is crucial here. Plesa-Skwerer, Verbalis, Schofield, Crawford, Ciciolla and Tager-Flusberg (2009) explored autonomic responsiveness to facial expressions of emotion. Whilst individuals with WS watched short dynamic clips of actors expressing 4 Eye contact and arousal in WS   emotional states galvanic skin response was recorded as a measure of autonomic responsiveness. Individuals with WS showed reduced arousal (hypo-arousal) in comparison to typically developing participants. The authors link this finding to the possible implications of atypical structure and functioning of the amygdala, given its role in mediating the autonomic nervous system. Interestingly, these results appear when individuals with WS are looking at images of faces on a computer screen, rather than engaging with another person. Therefore extrapolating this type of research to real-life interactions would be particularly beneficial and may reveal more about the implications of atypical amygdala functioning and its relationship to autonomic responsiveness in natural interactions. The second hypotheses for explaining to WS social phenotype is the involvement of atypical inhibitory control / attention processes (e.g. Porter, Coltheart, & Langdon, 2007, Cornish, Scerif & Karmiloff-Smith, 2007) which may not only allow faces to capture attention in an atypical manner, but may also relate to how attention is shifted or disengaged from socially relevant information. Currently however, the exact contribution of these factors remains unclear and indeed their contribution may not be mutually exclusive. Understanding not only the source of atypical social behaviour, but the exact nature of this behaviour is particularly important for understanding and appropriately teaching social skills to individuals with WS. Differences in the way that individuals with WS engage in social communication during face- to-face contact may explain difficulties with peer relations and associated social anxieties (e.g. Laws & Bishop, 2004), irrespective of the empathetic nature of individuals with the disorder (Gosch & Pankau, 1997). While face-to-face signals, like eye gaze and facial expressions, are beneficial to communication, they carry a cognitive load which typically developing children and adults 5 Eye contact and arousal in WS   avoid by averting their gaze (e.g. Doherty-Sneddon Bruce, Bonner, Longbotham, & Doyle, 2002). Furthermore making and holding direct eye contact with another person influences physiological arousal (linking to the work previously mentioned by Plesa-Skwerer and colleagues), leading to increased heart rate and skin conductance (e.g. Gale, Lucas, Nissim & Harpham, 1972, Kleinke, 1986, Andreassi, 2000, Argyle & Cook, 1976). Some researchers claim that one reason we avert our gaze is to reduce heightened physiological arousal caused by prolonged mutual gaze (e.g. Field, 1981). The first aim of the current research is to begin exploring the effect of direct eye contact on physiological arousal in a small group of individuals with WS. The previous work of Plesa-Skwerer and colleagues (2009) is particularly informative here but working with actual inter-personal situations would further advance our understanding of arousal and gaze behaviour in WS. The previous eye-tracking evidence has also emphasised the face gaze atypicalities shown by people with WS (Riby & Hancock, 2008, 2009), but again the inter-personal conditions created in the current research will extrapolate these findings to more naturalistic social interactions. It is hypothesised that individuals with WS will show less arousal to faces (hypothesis 1). In addition we predict that individuals with WS will avert their gaze less overall (hypothesis 2) and will not use GA in response to increasing cognitive difficulty of questions (hypothesis 3). Study 1: Face-to-face arousal effects Method Participants 6 Eye contact and arousal in WS   Fifteen individuals with Williams syndrome participated1, ranging from 8 years 10 months to 28 years 2 months (mean 14 years 11 month). All participants were recruited through the Williams syndrome Foundation. Parental and participant consent was received prior to testing. Eleven participants had previously had their diagnosis confirmed with the fluorescent in situ hybridisation diagnostic test (FISH), whilst the remaining four had been diagnosed phenotypically by clinicians. Three participants had to be removed from the study, one participant was unable to answer the mathematics questions and two were unable to remain still for the GSR measurements. Therefore the final sample comprised 12 participants (6 female, 6 male, 11 participants with positive FISH tests) aged 10 years 2 months to 28 years 2 months (mean 17 years 6 months). To compare physiological arousal to that of typically developing individuals, a comparison group of 12 participants was recruited and matched to each participant with WS on the basis of chronological age2. There was no difference in the chronological ages of the two groups (p=.48). The group with WS had a mean age of 17 years 7 months (as detailed above) whilst the typically developing group also had a mean age of 17 years 7 months (ranging from 10 years 4 months to 28 years 4 months). Stimuli and Procedure Participants completed 3 levels of task difficulty (no task; easy mental arithmetic- e.g. ‘count from 1-15 in 1s’; difficult mental arithmetic- e.g. ‘count backwards from 100 in 3s’) while 1 This small sample size is adequate in respect to other studies of individuals with this rare genetic disorder in the UK and for the aim of providing pilot data for future investigations of arousal in WS. The sample size is also comparable to studies investigating the gaze behaviour of children with other developmental disorders such as Fragile X syndrome (e.g. n=6, Hall et al., 2009). 2 In study 1 we are measuring physiological arousal which is documented in the literature as changing in relation to chronological age with younger children being more physiologically responsive than older children and adults (e.g. Venables & Mitchell, 1996). Hence it was important to have a chronological age matched control for this physiological study. 7 Eye contact and arousal in WS   looking towards the experimenter or at the floor (18 trials in total, order counterbalanced). Task difficulty was individually adjusted as necessary for each participant (see Paterson, Girelli, Butterworth, & Karmiloff-Smith (2006) on mathematics ability in WS). The calibration of difficulty was done a priori for all participants with WS in collaboration between the experimenter and the parents/guardians/teachers of the WS participants. However if a participant found a particular task unexpectedly challenging/easy this was altered or moved to a different category of difficulty as appropriate. For example, instead of counting backwards in 3s the participant may have counted backwards in steps of 2 or 1 depending on their individual capability. Physiological arousal was measured by recording changes in the participants’ skin conductance levels. Silver/silver chloride electrodes filled with gel were placed on the medial phalanges of the middle and index fingers of the participant’s hand and connected to a Biopac MP30 amplifier. The data were recorded and displayed in microSiemens (µSiemens). The mean skin conductance level (SCL) for each trial was calculated. After each trial, there was a rest period of 30-40 seconds in order that SCL levelled off before commencing the next trial. Results An ANOVA was conducted on the mean SCL data. Group (WS, TD), Condition (Face, Floor) and Difficulty (easy, difficult, no task) were the independent variables. Means are given in Table 1. There was a significant effect of Condition, F(1,22) = 21.03, p < .001, η 2 = p .49, with arousal significantly higher while looking at the face compared with the floor (mean face arousal = 6.78 , mean floor arousal = 6.03). There was a significant effect of Group, 8 Eye contact and arousal in WS   F(1,22) = 29.60, p < .001, η 2 = .57, as individuals with WS showed lower SCL levels than p those who developed typically (mean WS = 4.18, TD = 8.63). There was a significant effect of Difficulty, F(2,44) = 16.43, p <.001, η 2 =.43 (mean easy = 6.37; mean difficulty = 6.65; p mean no task = 6.20). Post hoc t-test revealed that participants had higher SCL scores when doing difficult tasks compared with both easy tasks and no tasks (t(47) = 3.25, p <.01; t(47) = 5.41 , p <.001). In addition participants had higher SCL when engaged on easy task compared to no tasks, t(47) = 2.20, p <.01. There was a trend to significance for the interaction between Group and Condition, F(1,22) = 3.83, p = .06, η 2 = .15. While this is a small effect size the p interaction has a direct bearing on our first hypothesis and hence planned comparisons t-test were used to explore it. These revealed that the effect of Group was significant in both the face and the floor conditions (t(22) = 5.75, p < .001, mean WS = 4.40, mean TD = 9.17; t(22) = 4.92, p < .001, mean WS = 3.96, mean TD = 8.10). Condition was significant for the typical developing participants (t(11) = 4.53, p <.001, mean face = 9.17; mean floor = 8.10) and there was a trend to significance for the WS participants (t(11) = 1.90, p = .08, mean face = 4.40; mean floor = 3.96). [ Insert Table 1 ] Discussion Individuals with WS were generally less physiologically aroused than age-matched typically developing individuals, with typically developing individuals showing around double the SCL. In addition both groups were more aroused when looking at faces (mean WS = 4.40; mean TD = 9.17) than the floor (mean W = 3.96; mean TD = 8.10), albeit the effect is marginal for the WS group. So contrary to our prediction faces were arousing for participants with WS although baseline levels of arousal were very low for this group in comparison to 9 Eye contact and arousal in WS   matched controls. The results may help explain (although there are likely many other factors involved) the anomaly in the WS literature that while people with WS are people orientated (Klein-Tasman & Mervis, 2003), relatively good at processing some aspects of faces (e.g. eye gaze cues, Riby et al., 2008b), and preferentially attend to faces for prolonged periods of time (Riby & Hancock, 2008, 2009; Mervis et al., 2003), they seem immune to the ‘overload’ associated with extensive eye contact. Using an assessment of physiological arousal it is possible to reveal that one of the important components of extended face gaze in WS is likely to be the effect of arousal level during social interactions and communication. Study 2: Gaze aversion while thinking Method Participants The WS group from study 1 took part. A new typically developing comparison group was recruited and individually matched on gender and verbal mental age to the participants with WS3. They had a mean chronological age of 9 years 4 months (ranging 5 years 4 months to 16 years 8 months). Verbal ability was assessed using the BPVS II (Dunn, Dunn, & Whetton, 1997), with participants with WS having a mean receptive vocabulary age of 8 years 11 months (ranging 4 years 7 months to 16 years 6 months) and typically developing participants 9 years 2 months (ranging 4 years 10 months to 16 years 11 months). There was no difference 3 The rationale for having a VMA matched control group in study 2 is because it examines gaze aversion behaviours during problem-solving. Abilities and strategies for problem solving are strongly linked to cognitive development and functioning, hence why a VMA match was taken. Furthermore GA increases with age (e.g. Doherty-Sneddon et al 2002) over the early primary years with most typically developing 8-year-olds behaving much like adults. Controls were around 8-years of age and we would therefore not expect them to differ much in terms of their GA behaviours from an adult chronologically matched control group. However because of the nature of the problem solving activity involved in study 2 we considered it important to have a mental age control. 10