Novel Behavioral Metrics of Attention Bias in Anxious Adolescents

• In Event: 3-035 - Rising From the Ashes? Paradigm Shifts in the Study of Attention Biases in Adolescent Anxiety

Integrative Statement

The anxiety-related attention bias (AB), or selective and exaggerated attention towards threatening information and stimuli, is thought to play a significant role in the etiology and maintenance of anxious pathology in children, adolescents and adults. Traditionally, AB has been quantified as the mean difference between reaction times (RTs) to threat versus neutral cues when both stimuli are competing for attention, thus providing a static snapshot of attention preference that obscures dynamic changes in AB over time. Indeed, a growing number of null and mixed findings have called into question the psychometric properties of these traditional AB metrics. For example, studies have shown low internal consistency, test-retest reliability, and cross-task convergent validity (Mogg et al., 2000; Van Bockstaele, 2014). Recently, the dynamic and context-sensitive characteristics of AB have been examined in studies of adult anxiety, and measures of temporal variability in AB have shown clinical significance, with higher variability predicting greater treatment response, but also higher levels of stress reactivity and anxiety severity (O’Toole & Dennis-Tiwary, 2018; Heeren et al., 2014; Zvielli et al., 2015). No studies to date have examined AB variability during adolescence, a developmental period with high rates of anxiety onset. The aim of the current study was to evaluate whether dynamic measures of AB (trial-level reaction times) predicted anxiety severity in a group of moderately to severely anxious adolescents (N= 18; 11 female; M age = 14.02, SD = 0.69). The Anxiety Disorders Interview Schedule (ADIS; DiNardo, Brown, & Barlow, 1994) was administered to both parent and child. Based on these interviews, a clinical severity rating (CSR) ranging from 1 to 8 was assigned. Anxiety symptoms were also evaluated using child self-report measures. Reaction-time metrics of AB, both the traditional average score and AB variability, were generated from the dot probe task. We tested the hypothesis that greater AB variability would predict higher levels of anxiety above and beyond variance accounted for by the average AB score. Hierarchical multiple regressions were conducted with the AB average score entered in the first step and AB variability in the second. AB variability, but not average AB, significantly predicted ADIS CSRs for primary diagnoses, Full model: R2 = .32, F(2,14) = 3.26, p = .069; AB variability: R2 = .32
[β = .55, t(15) = 2.54, p = .023]. AB variability, but not average AB, also predicted child-reported anxiety, Full model: R2 = .49, F(2,15) = 7.16, p = .01; AB variability: R2 = .49 [ β = .73, t(16) = 3.73, p = .002]; both average AB effects p >. 10. Results demonstrated that AB variability predicted anxiety symptoms and severity in a group of anxious adolescents, even when accounting for variance explained by average AB measure. Future research should continue to examine the clinical relevance of temporally-sensitive measures of AB, particularly during high risk developmental periods such as adolescence. Results represent an important step towards clarifying the conceptualization and measurement of AB and, ultimately, could speed progress towards more precise and targeted treatment approaches in anxious youth.

Authors

• Mariah Lillian DeSerisy, Fordham University
  Presenting Author

• Amy Krain Roy, Fordham University
  Non-Presenting Author