Abstract: This study employed a three-level meta-analysis to systematically investigate attentional biases in individuals with depression and the factors that may moderate these biases. A total of 51 studies on negative attentional bias (91 effect sizes) and 33 studies on positive attentional bias (56 effect sizes) were included. The findings revealed a moderate negative attentional bias together with a deficit in positive attentional bias. Moderator analyses showed that task paradigm significantly influenced both types of attentional bias, while stimulus type (pictures vs. words) had a marginal effect on negative bias and a significant effect on positive bias. In contrast, age, sample type, and stimulus valence did not exert significant moderating effects. These results highlight the dual features of attentional bias in depression and suggest that mood improvement may rely on both reducing excessive attention to negative information (“attenuation”) and enhancing attention to positive information (“augmentation”). This study provides a theoretical basis for refining cognitive models of depression and advancing dual-track intervention strategies.

Key words: depression, negative attention bias, positive attention bias, three-level meta-analysis