Background <p>Subclinical depression is a risk factor for the future onset of major depressive disorder, and a key target for intervention. Computerised interventions are increasingly being employed to manage subclinical depression, though their effectiveness and mechanisms of action are unclear. Self-help style CBT techniques such as cognitive restructuring, and cognitive training targeting executive function/working memory are two possible computerised interventions. This study investigated the effectiveness of working memory training and cognitive restructuring at reducing depressive symptoms, and whether changes in interpretative bias and working memory mediated this improvement.</p> Methods <p>104 adults (67 female) with mild to moderate depression and anxiety were randomly assigned to either a waitlist control group, or to complete either a working memory training, or a cognitive restructuring training. Depressive symptoms, interpretive bias and working memory were assessed pre- and post-training.</p> Results <p>Cognitive restructuring reduced depressive symptoms and improved interpretative biases post-training, though exploratory analyses suggested this may have been due to a regression to the mean, and there was no effect of working memory training on either outcome (although very few participants adhered to the protocol/recommended training frequency). Improvements in interpretative bias and executive function did not mediate training-related changes in symptoms.</p> Conclusions <p>Cognitive restructuring reduced depressive symptoms after a 2-week training period, suggesting regular practice may be effective in a subclinical population. However, working memory training does not appear acceptable or effective unless delivered under highly controlled conditions to ensure adherence. Additionally, we found no effect of training on either putative mechanism of action, suggesting further research is required.</p>