Williams syndrome (WS) is a sporadically occurring, relatively rare, developmental disorder caused by a hemizygous deletion of approximately 25–28 genes on chromosome 7q11.23 (Hillier et al.
2003). It has a reported prevalence of 1:20,000 (Korenberg et al.
2003, but see also 1:7,500; Strømme et al.
2002), and affects males and females equally. WS is associated with a distinctive profile of medical, physical, cognitive and behavioural characteristics. For example, individuals with WS can have medical difficulties including heart problems (supravalvular aortic stenosis), hypercalcemia, musculoskeletal abnormalities, and distinctive facial morphology (Morris
2006).While there is considerable variability of intellectual functioning, most individuals with WS have mild to moderate cognitive impairments (Mervis et al.
2000). Behaviourally, individuals with WS tend to be very friendly and empathetic, and have often been described as hypersociable (Jones et al.
2000). In more recent years there has been a focus on the social atypicalities (Lough et al.
2015) and psychiatric comorbidities, such as anxiety (Dykens
2003). Anxiety is the most prevalent mental health concern, especially by adulthood (e.g. Stinton et al.
2012). However, despite this heightened prevalence, interventions targeting anxiety are lacking (e.g. Cherniske et al.
2004). As future intervention strategies require a comprehensive understanding of anxiety in WS, the present study focuses on the phenomenology, development, and correlates of this phenomenon. It is important to consider how anxiety might be associated with the other characteristics of the disorder, specifically aspects of the combined cognitive and social phenotypes.
Anxiety in Williams Syndrome
Alongside a socially gregarious disposition, the high prevalence of anxiety-related psychopathology is a seemingly paradoxical feature of WS. In a comprehensive study of anxiety in WS, Leyfer et al. (
2006) assessed the occurrence of co-morbid psychiatric disorders in 4–16 year-olds with WS (N = 119). Using the Anxiety Disorder Interview Schedule (Silverman and Albano
1996), a structured clinical interview for parents, they found that rates of generalised anxiety disorder (GAD; 12%) and specific phobia (SP; 54%) were significantly higher than observed in the general population and those with intellectual disability. Furthermore, rates of GAD were higher among older individuals (11–16 years) than would be expected based on rates in younger individuals (4–6 years), suggesting anxiety in WS increases with age. Elsewhere, Cherniske et al. (
2004) assessed the psychiatric profiles of 20 adults with WS. Based on diagnostic assessments by licensed psychiatrists, 13 individuals were classified as having moderate or severe anxiety, while three were described as having milder, subclinical problems. Again, the most common anxiety disorders were SP and GAD. By extending the previous findings to an adult sample, this study further supported the claim that anxiety is a persistent phenomenon in WS.
The studies outlined above employed cross-sectional designs, preventing inferences about the development of anxiety over time. To date, there are two known longitudinal studies of anxiety in WS. First, Woodruff-Borden et al. (
2010) used the Anxiety Disorder Interview Schedule to assess 4–13 year-old children with WS (N = 45; mean age 6.67 years) in a two-year longitudinal study. While 60% of their sample presented with at least one anxiety disorder on initial assessment, this figure had increased to over 80% by follow-up. Seventy-two per cent of those with an initial anxiety disorder had developed an additional diagnosis at follow-up. Thus, the frequency of additional cumulative diagnoses suggests that in WS, anxiety generally remains stable, and in some individuals increases over time.
Not all findings support this trend though, as Green et al. (
2012) conducted a five-year longitudinal study, exploring rates of psychiatric disorders among 6–23 year-old children with WS (N = 38) and developmental disabilities of mixed etiology. Using the Kiddie Schedule for Affective Disorders and Schizophrenia (Kaufman et al.
1997), they found that rates of anxiety disorders were significantly higher in WS at both time points. However, the WS group also showed a dramatic decrease in prevalence rates from initial assessment (84%) to follow-up (44%). A small proportion of the WS children in this sample received SSRI medication over the study period, which may have influenced outcomes. Furthermore, the average age of the sample at initial assessment was 13 years, as opposed to 6 years in Woodruff-Borden et al. (
2010) study, and it may be that the trajectory of anxiety changes from childhood to adolescence. Given these differing accounts, further longitudinal studies across a wider age range are warranted to clarify the developmental trajectory of anxiety in WS.
Anxiety and Social Functioning in Williams Syndrome
A defining feature of WS is the social profile, characterized by heightened attraction to faces and a significant motivation towards social interactions (Jones et al.
2000; Frigerio et al.
2006). Numerous reports describe people with WS as “hypersocial” compared to those with other developmental disorders and typically developing individuals (see Järvinen-Pasley et al.
2008). However, despite strong affiliative tendencies, individuals with WS consistently score within the range of mild-to-moderate impairment on social reciprocity measures (e.g. Kirk et al.
2013; Lough et al.
2015). These difficulties have lasting impacts on adaptive functioning and well-being. For example, undiscerning social approach behaviours—such as indiscriminately engaging others without considering social cues—coupled with cognitive impairments can increase the potential for victimization and social vulnerability (Klein-Tasman et al.
2011; Lough et al.
2015; Jawaid et al.
2012; Riby et al.
2017). Furthermore, many adults with WS are unable to maintain friendships, and suffer from social isolation (Howlin and Udwin
2006).
Attempts to understand the socio-emotional and behavioural correlates of anxiety in WS have led to an emerging body of research into the potential interplay between anxiety and social functioning. Riby et al. (
2014) measured parental reports of anxiety using the Spence Children’s Anxiety Scale (SCAS; Spence
1998)—and social functioning—using the Social Responsiveness Scale (SRS; Constantino and Gruber
2005) in their study of 59 individuals with WS aged between 6 and 36 years. They found a small positive but significant correlation between anxiety and social functioning impairments (r = .362,
p < .01). Splitting the participants into high- and low- anxious groups revealed that highly anxious individuals had greater impairments on the SRS subscales social awareness, cognition, and communication, which are said to reflect “socio-cognitive” functions (Klein-Tasman et al.
2011). Interestingly, social motivation did not differ between the groups. In other words, although both groups were similarly motivated by social interactions, the high-anxious individuals were less adept in socio-cognitive domains. This does not support the idea that social motivation serves as a protective factor against anxiety but rather that hypersociability may mask anxiety in social situations (Dodd et al.
2009; Dykens
2003). Therefore, when trying to understand anxiety in WS it is important to account for the role of social functioning.
Anxiety and Executive Functioning in Williams Syndrome
Executive functions (EFs) have become a topic of increased focus within the WS cognitive profile. They are a group of higher-order cognitive processes associated with pre-frontal circuits believed to modulate cognitive, social, and emotional behaviours. They are widely conceptualized as a set of separate but related constructs, including cognitive flexibility (or shifting), inhibition, and working memory (Miyake et al.
2000). Several studies have found delays and impairments across a range of EFs in WS; including inhibition, set-shifting, and working memory (e.g. Menghini et al.
2010; Rhodes et al.
2010). Furthermore, while some developmental improvements in EF are observed during early childhood, deficits generally persist into adulthood (Greer et al.
2013).
Evidence from typical development indicates that executive dysfunction is associated with higher trait anxiety (Ursache and Raver
2014), and moderates the relationship between having an anxious temperament and developing an anxiety disorder (Fox
2010). In a study of adults with WS (N = 19) Rhodes et al. (
2010) administered a battery of tasks measuring attention set-shifting, planning, and working memory abilities. They found that executive dysfunction across these tasks was associated with parental reports of negative affect, conduct problems, and decreased prosocial behaviours, as measured by the Conners Parent Rating Scale (Conners et al.
1998), and the Strengths and Difficulties Questionnaire (Goodman
2001).
Elsewhere, McGrath et al. (
2016) explored the relationship between anxiety and attentional control in individuals with WS (aged 12–56), using the SCAS and a social dot-probe task in which participants (N = 46) were exposed to either happy or angry faces. They reported that highly anxious individuals displayed an increased bias towards angry faces, which was primarily explained by an inability to
disengage, or shift attention from threatening stimuli. They noted that as general “sticky attention” effects are well-documented in WS (e.g. Riby and Hancock
2008; Riby et al.
2011).It is possible that a broader difficulty in attention-shifting, coupled with hypervigilance to threat, underlies the onset and maintenance of anxiety in this population.
In response to suggestions that performance-based EF tasks are too reductionist (e.g. Brown
2006), rating scales such as the behaviour rating inventory of executive functioning (BRIEF; Gioia et al.
2000) have been designed to measure EFs with greater ecological validity, by allowing researchers to assess regulatory abilities in everyday settings (Kenworthy et al.
2008). So far, the BRIEF has yielded several important findings in WS research. For example, Woodruff-Borden et al. (
2010) found that the presence of an anxiety disorder was associated with increased behavioural and emotional dysregulation. Moreover, this relationship was stable over time. As the authors only reported difficulties in broader domains of behavior and emotion regulation, it would be interesting to build on these findings, honing in on the specific functions associated with increased anxiety.
More recently, Pitts et al. (
2016) investigated the association between Specific Phobia (SP—as measured by the ADIS-P) and behavioural regulation (as measured by the BRIEF) in their sample of children and adolescents with WS (N = 194). Using a logistic regression model, they found that behavioural dysregulation was the strongest predictor of SP, with children at or above clinical levels of dysregulation at the greatest risk for SP. The authors proposed that impaired abilities to self-regulate or shift attention away from threatening stimuli leads to the subsequent development of irrational fears surrounding specific objects or situations. While this study has important implications for understanding some of the cognitive and behavioural foundations of anxiety in WS, its scope is limited as it only covers the associations between EF and SP. As such, it is unclear whether behavioural dysregulation confers an increased risk for both cue-specific and more generalized forms of anxiety (such as GAD). It is thus important to build upon this work by investigating whether the reported relationship extends beyond cue-specific anxiety disorders towards a broader range of anxious symptomatology.
Therefore, although the evidence suggests that EF impairments are implicated in the presence of anxiety in WS, further work is needed to address gaps in current understanding. For example, questions remain as to the relative contribution of different components of EF to the development and maintenance of anxiety in WS, and the associations to other key characteristics of WS, namely social functioning.
The Current Study
In light of the above literature, the aims of this study were two-fold. The first aim was to explore the developmental course of anxiety in WS by looking at changes in its presentation over time (4 years) and its association with age (in the cross-sectional sample). Based on existing literature, anxiety was predicted to increase over time, and with age. A second aim was to explore how anxiety was associated with other core features of the disorder, namely social and executive functioning. It was predicted that most individuals with WS would show impairments in both social and executive functioning. Specifically, those with higher levels of anxiety were predicted to present with more difficulties with social functioning. We also explored whether specific aspects of executive functioning (i.e. shift, inhibit, etc.) were differentially associated with anxiety, although due to limited literature no specific predictions were made. Finally, the study sought to examine the extent to which social and executive functioning predicted anxiety in the cross-sectional sample. As no studies to date have measured all three aspects of the WS psychosocial profile concurrently, no specific hypotheses were posed.