Attention Alterations in Pediatric Anxiety: Evidence From Behavior and Neuroimaging

Abstract

Background: Pediatric anxiety disorders involve greater capture of attention by threatening stimuli. However, it is not known if disturbances extend to nonthreatening stimuli, as part of a pervasive disturbance in attention-related brain systems. We hypothesized that pediatric anxiety involves greater capture of attention by salient, nonemotional stimuli, coupled with greater activity in the portion of the inferior frontal gyrus (IFG) specific to the ventral attention network (VAN).

Methods: A sample of children (n = 129, 75 girls, mean 10.6 years of age), approximately half of whom met criteria for a current anxiety disorder, completed a task measuring involuntary capture of attention by nonemotional (square boxes) and emotional (angry and neutral faces) stimuli. A subset (n = 61) completed a task variant during functional magnetic resonance imaging. A priori analyses examined activity in functional brain areas within the right IFG, supplemented by a whole-brain, exploratory analysis.

Results: Higher clinician-rated anxiety was associated with greater capture of attention by nonemotional, salient stimuli (F_1,125 = 4.94, p = .028) and greater activity in the portion of the IFG specific to the VAN (F_1,57 = 10.311, p = .002). Whole-brain analyses confirmed that the effect of anxiety during capture of attention was most pronounced in the VAN portion of the IFG, along with additional areas of the VAN and the default mode network.

Conclusions: The pathophysiology of pediatric anxiety appears to involve greater capture of attention to salient stimuli, as well as greater activity in attention-related brain networks. These results provide novel behavioral and brain-based targets for treatment of pediatric anxiety disorders.

Keywords: Attention; Inferior frontal gyrus; Network neuroscience; Pediatric anxiety; Ventral attention; fMRI.

Figure 1.

The task used to measure involuntary capture of attention. To begin each trial, 1 of 4 cue types appeared randomly at the left or right side of the screen for 150 ms. After a random delay of 50, 350, or 650 ms (resulting in a cue-target onset asynchrony of 200, 500, or 800 ms), a target appeared randomly on the left or side of the screen. The participants’ task was to indicate whether the target was pointing upward or downward by pressing a button on a keyboard. Measures of attention are calculated as how much faster individuals were when the cue was on the same side of the screen as the target (valid trials) relative to trials when the cue was on the opposite side of the screen (invalid trials).

Figure 2.

Anxiety is positively related to involuntary capture of attention by nonemotional salient stimuli. (A) The specific trial type that measures the initial capture of attention by salient, nonemotional cues. (B) The magnitude of the initial capture of attention by nonemotional cues in children with high versus low anxiety. This magnitude is calculated as how much faster reaction time was for validly cued trials with the 200-ms cue-target onset delay relative to invalidly cued trials. (C, D) Reaction times for participants with low/high anxiety (median split of Pediatric Anxiety Rating Scale) for discriminating targets appearing at the same (valid trials) or opposite side (invalid trials) of the screen relative to the square-box cue in trials with the shortest cue-target delay. All statistical analyses used anxiety scores (Pediatric Anxiety Rating Scale) as a continuous variable.

Figure 3.

Significant activity change (p < .05) associated with the task. The task elicits widespread brain activity, including in brain networks implicated in executive function and control of attention. The images above illustrate the main effect of task in the imaging analysis, similar to average evoked activity regardless of level of anxiety. Warm colors indicate activity increases, while cool colors indicate activity decreases relative to a fixation baseline. Borders indicate the boundaries of functional brain networks as determined by a study of adults (70). Activity modulations are noted across much of the brain, including in the ventral attention (Vent Attn), dorsal attention (Dors Attn), salience, frontoparietal (Fronto-Par), and cingulo-opercular (Cing-Operc) networks. Note that activity in the motor network was restricted to the hand representation in the left hemisphere, consistent with subjects making right-handed button presses. Par, parietal; SM, somatosensory-motor.

Figure 4.

A priori and exploratory analyses examining activity change associated with anxiety. (A) The 8 a priori defined functional areas within the inferior frontal gyrus (IFG) in which we explored the relation between anxiety and brain activity during the attention task. The outlines depict all regions, while the solidly colored area is the portion of the IFG within the ventral attention network (VAN), which was the only region in which activity was significantly related to anxiety after Bonferroni correction. (B) Brain activity elicited in this IFG region within the VAN during the involuntary capture of attention is correlated with anxiety severity (Pediatric Anxiety Rating Scale [PARS]), while controlling for age and sex. Results were unchanged when excluding the single outlier data point. (C) Results from our exploratory whole-brain analyses, examining the effect of anxiety during the attention task, while controlling for age and sex. Borders indicate the boundaries of functional brain networks as determined by a study of adults (70). Using a threshold of p ≤ .01 and surface area ≥100 mm², we observed 3 clusters, (A) in the left superior temporal gyrus (L STG), (B) near the a priori selected right IFG (R IFG), and (C) in the R frontal pole (default mode network). In each case, higher anxiety was associated with higher regional brain activity. BOLD, blood oxygen level–dependent. Attn, attention; Cing-Operc, cingulo-opercular; Dors, dorsal; Fronto-Par, frontoparietal; Par, parietal; SM, somatosensory-motor; Vent, ventral.