Sensitivity and Bias in Face-Emotion Labeling: Replication and Extension to Youth With Irritability and Anxiety

Pediatric anxiety and irritability are common, impairing, co-occurring symptoms. Biases in interpreting ambiguous face-emotions have been linked to both phenotypes. Here, we assessed whether biases represent a shared cognitive and neural profile. In addition, we attempted replication of prior age associations and group behavioral differences in face-emotion labeling.

In a cross-sectional functional magnetic resonance imaging study at a research facility at the National Institute of Mental Health, we used a variant of a drift diffusion model to decompose perceptual and cognitive components of binary happy-angry decisions to ambiguous face-emotions. A total of 95 participants (mean age = 14.11, SD = 3.07, range = 8-22, male = 50.9%) contributed to the analyses. In all, 65 participants did not complete the scan or meet data quality thresholds. Participants had a primary diagnosis of disruptive mood dysregulation disorder (DMDD; n = 27), attention-deficit/hyperactivity disorder (n = 23), oppositional defiant disorder (n = 2), anxiety disorder (n = 19), or no current diagnosis (n = 24), resulting in a wide range of anxiety and irritability symptom levels.

No significant associations between the computational modeling parameter indexing bias and dimensionally assessed anxiety or irritability emerged (r[84] < -0.15, p > .05). Parent- and child-rated irritability, but not anxiety, was associated with increased neural responses to overtly angry faces, most notably in the motor cortex (voxel-wise p < .005). We successfully replicated prior age-associated increases in sensitivity and differences in bias between youth with DMDD and controls.

Increased neural engagement of the bilateral motor cortex likely reflects increased vigor in responding to overtly angry faces, possibly indexing approach behavior. Replication of prior findings increases confidence in the robustness of these associations.