Abstract
BACKGROUND: Depressive (DD) and bipolar (BD) disorders are characterized by biases in anticipating and encoding emotional information. Depressive traits are linked to negative affective biases, while hypomanic features are associated with heightened responsiveness to positive stimuli. The vividness of visual imagery may further modulate these biases. This study examined how lifetime dimensional symptoms of depression and hypomania across diagnoses interact with imagery vividness to modulate brain activation during anticipation and encoding of happy and sad faces. METHODS: A total of 155 individuals aged 18-45 years with BD, DD, or healthy control (HC) status completed a cued emotional face-encoding task during functional magnetic resonance imaging (fMRI). Lifetime dimensional symptoms of depression and hypomania were assessed using the MOODS-SR, and imagery vividness was measured with the Vividness of Visual Imagery Questionnaire (VVIQ). Interaction effects between spectrum depression/hypomania and imagery vividness on brain activation during anticipation and encoding of happy versus sad faces were analyzed using the Sandwich Estimator approach in FSL. RESULTS: Higher spectrum hypomania scores were associated with greater imagery vividness and increased activation for happy versus sad faces in the occipital pole, cuneus, intracalcarine, and lateral occipital cortices during encoding. However, there was reduced activation for happy versus sad faces in the precentral gyrus during anticipation. Depressive symptoms interacted with imagery vividness within the default mode network: more severe spectrum depression and lower vividness were associated with greater activation for happy versus sad faces in the frontopolar cortex during anticipation, but with reduced activation in the angular gyrus during encoding. CONCLUSION: Lifetime depressive and hypomanic spectrum symptoms across diagnoses differentially interact with visual imagery to influence anticipation and encoding of emotional faces. Depressive biases were observed in frontopolar-parietal regions, while hypomanic biases appeared in occipital cortices. These findings highlight the value of dimensional approaches to mood psychopathology and identify imagery vividness as a promising transdiagnostic target for intervention.