(PS3-D88) Differences in Attentional Threat Bias Among Youth with Low vs. High Clinician-rated Anxiety

Childhood anxiety is characterized by attentional threat bias, or hypervigilance towards threatening information. Traditional assessment of pediatric anxiety and threat bias have often relied respectively on biased self-report measures and limited response time paradigms. This study assesses anxiety through clinician ratings, considered the “gold standard” of symptom measurement, and threat bias through a continuous eye-tracking paradigm. To better understand the anxiety-threat bias relationship, we compared individuals with high and low clinician-rated anxiety on a behavioral threat bias task.
A transdiagnostic sample of youth (N = 153, M age = 12.9 +/- 2.8 years, 28.8% female, 64.1% with anxiety, mood, or attention disorders, 35.9% healthy controls) viewed matrices of faces with contrasting emotional valences during an eye-tracking task. Sustained attention allocation (i.e., “dwell time”) was captured separately for time spent viewing Negative and Non-Negative facial valences. Dwell bias scores were calculated by subtracting Negative - Non-Negative dwell times. Child anxiety symptoms were measured by the Pediatric Anxiety Rating Score (PARS), a clinician-rated semi-structured interview. The PARS includes 7 subscales, and total PARS score included a sum of 5 subscales. Due to abnormally skewed data, continuous PARS variables were dichotomized into low vs. high groups. Independent samples t-tests were conducted to examine differences in dwell times and dwell bias scores between high and low PARS total and subscale groups.
There were no significant differences in dwell times or dwell bias scores between low vs. high PARS total score groups (ps >.05). There were no significant differences in dwell times or dwell bias scores across low vs. high groups of PARS subscales assessing frequency of anxiety symptoms, severity of distress associated with symptoms, severity of physical symptoms, and avoidance (ps >.05). There was a significant difference in Non-Negative dwell time between the low (M = 5792.14, SD = 1892.97) and high PARS anxiety symptom number group (M = 4890.49, SD = 22118.65), t(151) = 2.51, p = 0.01. There was a significant difference in dwell bias scores between the low (M = -101.48, SD = 1232.19) and high interference with family group (M = 389.22, SD = 1074.44); t(151) = -2.62, p = .01, and the low (M = -17.07, SD = 1217.06) and high interference with peer/adult group (M = 347.59, SD = 1100.74); t(151) = -1.95, p = .05.
Dwell times and dwell bias scores did not differ across low vs. high PARS total score groups but did differ across subscales measuring high vs. low number of anxiety symptoms, interference with family relationships, and interference with peer/adult relationships. Clinicians may use number of anxiety symptoms, rather than a global anxiety rating, to identify youth at highest risk for exacerbated threat bias. Interventions that reduce threat bias may be particularly relevant for youth who are experiencing high interference in their social relationships due to anxiety. These findings encourage continued research into the nuances of the threat bias-anxiety relationship and highlight threat bias as a potentially clinically relevant target in treatments of anxiety-based relationship interference.