Decision dynamics of negative bias in subclinical depression during ambiguous facial emotion perception

Abstract Individuals with depressive tendencies may categorize ambiguous facial expressions toward negative rather than happy endpoints, but the associated decision-process components remained unclear. The study examined whether the categorization shift was associated with differences in perceptual criterion, discrimination sensitivity, response dynamics, or diffusion-model parameters. Fifty-nine participants, comprising a depressive tendency group and a comparison group, categorized morphed Sad–Happy and Angry–Happy expressions presented in a 750-ms face/mask sequence or a 1500-ms face/no-mask sequence. Psychometric functions, area-under-the-curve (AUC) analyses, quadratic mixed-effects models of response time, and a hierarchical drift–diffusion model were applied. Across conditions, the depressive tendency group showed lower points of subjective equality than the comparison group, indicating a relative shift in the categorization boundary toward the negative versus happy endpoint, without a corresponding group difference in discrimination sensitivity. AUC analyses similarly indicated higher negative-response probabilities in the depressive tendency group, particularly within the low-negativity portion of the continuum. Condition-specific patterns were observed, but direct tests did not support a larger Sad–Happy group difference in the 1500-ms than in the 750-ms presentation sequence. In the focal Sad–Happy 1500-ms face/no-mask sequence, the selected drift–diffusion model indicated a group-related drift-criterion shift toward the Sad boundary and modestly longer non-decision time in the depressive tendency group, whereas starting-point bias and boundary separation did not differ between groups. These findings are consistent with a relative categorization shift toward negative versus Happy facial-expression endpoints among participants with depressive tendencies. The diffusion-model results provided a model-dependent account of the focal choice–response-time pattern under the present presentation-sequence structure.